![Chapter 1
Motivation and goals
This chapter provides a historical review of satellite communications, in partic-
ular from the perspective of the multiple access protocols. From this review, the
motivation and goals for the present PhD thesis are derived.
1.1 Motivation
1.1.1 A brief history of satellite communications and services
The Space Race (1955-1972) between the Soviet Union (USRR) and the United
States (US) opened up the era of satellite communications. In 1957, the Soviet
Union launched the first communications satellite (Sputnik 1) into an elliptical
Low Earth Orbit (LEO). It was a 58 cm diameter polished metal sphere, with
four external radio antennas to broadcast radio pulses. In 1958 the US launched
the first satellite which broadcast a recorded Christmas greeting from President
Eisenhower during 13 days (Project SCORE). In 1960 the US launched the first
active repeater communications satellite (Courier 1B) and was also the first
communications satellite powered by solar cells to recharge storage batteries
[1]. In 1962 the US launched Telstar 1 and Relay 1 wideband (microwaves)
repeater satellites on elliptical LEO orbits, which relayed the first live television
signals across the atlantic and pacific oceans respectively. Syncom 3 was the first
geostationary (GEO) communication satellite, launched on 1964, and was used
to broadcast the 1964 Summer Olympics in Tokyo to the United States. The
first commercial geosynchronous satellite, Intelsat I, was launched in 1965, as
well as, the first Soviet communication satellite on an inclined Highly Elliptical
Orbit (HEO), Molniya 1. Such orbits allowed the satellites to remain visible to
sites in polar regions for extended periods, unlike geostationary satellites which
appear below the horizon at very high latitudes. In 1974, NASA launched the
ATS-6 satellite, which was a precursor to many technologies still in use today on
geostationary spacecraft: large deployable antenna (9 m), direct to home (DTH)
broadcasting, 3-axis stabilized and antenna pointing through RF sensing. The
first in a series of Soviet GEO satellites to carry DTH television was launched on
26 October 1976 (Ekran 1). They were both using ultra-high frequency (UHF)
band (around 700-800 MHz), so that the transmissions could be received with
3](https://image.slidesharecdn.com/resolutionforrandomaccessschemes-220518201640-39d36c53/85/Resolution-for-Random-Access-Schemes-pdf-17-320.jpg)
![4 CHAPTER 1. MOTIVATION AND GOALS
existing UHF television technology. Due to the limited performance of the first
satellites, initial systems had low capacity, required expensive large earth stations
with antennae of 15 to 30 m diameter and were used to gather the traffic from
an extensive area by means of a ground network and to relay the signals across
the oceans (point-to-point trunk connections).
During the 1980s and the 1990s, the development of satellite microwave
(GHz) technologies, such as contoured multi-beam antennae with higher gain
and high power transmission amplifiers (200-300W), enabled the use of smaller
earth stations affordable to the end user. These smaller earth stations, typi-
cally have a satellite dish antenna ranging from 0.6 to 3.5 m diameter and can
be receive-only DTH terminals or bidirectional Very Small Aperture Terminals
(VSAT). Their smaller size also enabled mobile services (e.g. on vessels, trains,
airplanes). Initial VSAT terminals supported download speeds ranging from few
hundred kbps to tens of Mbps and upload speeds ranging from few kbps to few
Mbps. The reduced size of satellite terminals resulted in a rapid increase in the
number of VSAT networks (including networks with several hundreds of sites),
and the emergence of new narrowband and broadband satellite communications
services exploiting the satellite’s natural capability to collect or broadcast signals
from or to several locations over a very large coverage area: corporate networks,
multi-point data transmission networks (retail point-of-sale or automatic teller
machine transactions), data collection networks (polling RFID tags, supervisory
controlled and data acquisition - SCADA), satellite news gathering, satellite
broadcast networks (to cable heads or directly to the consumer) [2]. These satel-
lites where typically operating in the C-band (4-8 GHz) and Ku-band (12-18
GHz).
Nowadays, most communications GEO satellites carry dozens of transpon-
ders, each with a bandwidth of tens of megahertz, providing a very high capac-
ity at a reduced communications cost. Communications satellites have a typical
launch mass between 3000 and 6000 kg, a payload DC power between 5 kW and
20 kW and a design lifetime of 15 years. Most transponders operate like a ”bent
pipe”, relaying back to earth the received signals after performing amplifica-
tion and a shift from uplink to downlink frequency. High Throughput Satellites
(HTS), such as Anik F2 (2004), Thaicom 4 (2005), Spaceway 3 (2007), KA-SAT
(2010), ViaSat-1 (2011), Yahsat Y1B (2012), EchoStar XVII (2012), Inmarsat
Global Xpress (2013, 2015, 2016) and Intelsat EpicNG (2016), offer capacities
around 100 Gbps which significantly reduces the cost per bit. The majority
of HTS operate in Ka-band (20-30 GHz), and the massive increase in capacity
is achieved by employing frequency re-use across multiple narrow spot beams
(usually in the order of 100s of kilometers per spot beam) [3]. ViaSat -1 satel-
lite has 72 spot beams, KA-SAT has 82, EchoStar XVII has 60 and Inmarsat 5
has 89 spot beams. These satellites support two-way communications at higher
speeds (e.g. 50 Mbps download and 10 Mbps upload) with small satellite dish
antennae (60-70 cm). Typical services offered are broadband connectivity to the
consumer and professional markets both fixed and mobile (e.g. land mobile, mar-
itime, aeronautical). The Alphasat communications satellite (2013) is another
example of a large spacecraft, operating in L-band (1-2 GHz) with a launch mass](https://image.slidesharecdn.com/resolutionforrandomaccessschemes-220518201640-39d36c53/85/Resolution-for-Random-Access-Schemes-pdf-18-320.jpg)
![1.1. MOTIVATION 5
of 6650 kg, a payload DC power of 12 kW and a a 12-meter aperture antenna
reflector. L-band satellites are well suited for mobile satellite services requiring
low data rates, low cost and small size terminals (e.g. handheld, machine to
machine communications).
Today, we have around 600 operating communications satellites in orbit,
with orders for 20-25 new commercial GEO satellites each year [4]. In 2013
there were about 200 million satellite TV subscribers worldwide (DTH service).
These are served primarily via C-band and Ku-band broadcast satellites. To-
day they are offering Standard Definition (SD) and High Definition (HD) TV.
In the near future they will also serve Ultra HD TV and interactive TV ser-
vices. Satellite TV services account for 78% of all satellite services revenues.
Regarding the consumer broadband access, there were 2.5 million subscribers
(mostly in North America) in 2014 and is expected to grow worldwide to 8.8
million subscribers in 2023, according to Euroconsult forecasts [5]. Enterprise
data services are also gradually migrating from broadcast to HTS satellites due
to the lower cost per bit, in particular VSAT networks and cellular (3G and
4G) backhaul. Furthermore, it is also foreseen that HTS will also see capac-
ity demands coming for commercial mobility (e.g. commercial ships, business
jets, commercial airlines). Analysts from Northern Sky Research (NSR) forecast
that High Throughput Satellites capacity will increase from 500 Gbps in 2013
to 2 Tbps in 2023 [6]. Another market with a strong growth via satellite are
Machine-to-Machine (M2M) communications (e.g. remote control and monitor-
ing in transportation and cargo, oil and gas, utilities, military, civil government
markets) [7]. Key strengths for M2M via satellite are their reliable connectivity
in remote and underserved regions, their global coverage and their independence
from terrestrial networks for mission critical applications. The total market fore-
cast by NSR in 2023 for satellite M2M services is $ 2.5 billion with 6 million
terminals sold. Majority of the market will be served in L-band, being partic-
ularly well suited for this type of services with lower cost and smaller physical
size terminals.
1.1.2 The multiple access problem
In the early years of satellite communications, the services provided were tele-
phone and television signals transmission between continents (point-to-point
trunk connections). The transmitted signals where analog and each carrier con-
veyed either a single TV signal or frequency division multiplexed telephone chan-
nels [8]. Multiple access to the satellite was done by fixed Frequency Division
Multiple Access (FDMA) assignments to each earth station.
For Direct To Home (DTH) television, receive-only terminals are used. The
television signals transmitted by the broadcast centre (point-to-multipoint) were
initially analog using the NTSC, PAL or SECAM standards and Frequency Mod-
ulation (FM). Then, in the mid-1990s digital television standards where adopted
exploiting the flexibility to multiplex several TV channels in one carrier and a
reduced bandwidth occupation [9], [78]. The digital television signals are trans-
mitted by the broadcast centre using a time division multiplex/phase shift keying](https://image.slidesharecdn.com/resolutionforrandomaccessschemes-220518201640-39d36c53/85/Resolution-for-Random-Access-Schemes-pdf-19-320.jpg)
![6 CHAPTER 1. MOTIVATION AND GOALS
modulation (TDM/PSK).
VSAT networks use bidirectional communications and typically have a star
topology where VSATs communicate to a larger master earth station (called
hub), but may also communicate directly to other VSATs when in mesh config-
uration [11]. The transmitted signals are digital using a PSK modulation. The
use of digital transmission offers a greater flexibility to multiplex different types
of services (television, voice, data) over the same satellite channel. Digital trans-
mission also allows an easier interworking with the digital terrestrial networks.
A single outbound carrier provides data transfer from the hub to the remote
terminals using a TDM/PSK waveform as for DTH television. The inbound
channel (multipoint-to-point) typically consists of several lower bit rate carriers
operating in Multi-Frequency Time Division Multiple Access (MF-TDMA).
The arrival of the VSAT networks, triggered the need for new multiple access
protocols able to efficiently share the satellite communication resources among
all VSAT terminals while maintaining acceptable throughout, loss and delay
performance. Multiple access is one of the most critical elements to the perfor-
mance of VSAT networks. Their design is often constrained by their operational
environment. The satellite environment is characterized by radio link impair-
ments and a large propagation delay, amongst other inherent attributes (e.g.
non-linearities). Typical radio link impairments in the satellite channel are fad-
ing and multipath interference. The large propagation delay, in the range of
250 ms for a geostationary satellite, represents a very particular property of the
satellite environment that will condition the applicability of terrestrial multi-
ple access schemes over the satellite environment. In the satellite channel, the
propagation delay is much larger than the time taken to transmit a packet, and
a sender may have sent several packets before the receiver starts receiving the
first packet. Satellite multiple access schemes must be able to deal with all these
characteristics.
During the 1980’s and 1990’s numerous multiple access protocols were devel-
oped [12, 13, 14, 15, 16, 17, 18, 19]. The available inbound bandwidth resources
are divided using FDMA, TDMA, MF-TDMA or Code Division Multiple Access
(CDMA) [8]. But as opposed to the fixed assignment multiple access techniques,
the new proposed protocols combined demand assignment, random access and
fixed assignment to dynamically control the access to the shared resources by
the many contending users. These protocols typically have a centralized control
(e.g. in the hub) which is managing the access to the satellite resources, as it
allows a more efficient multiplexing of services with different priorities (e.g. data,
voice, video).
First, with the telephone service a portion of the satellite resources (e.g. fre-
quency channel in FDMA or timeslots in TDMA) were assigned dynamically
on a call basis [20] and the grade of service was determined through the the
well-known blocking formulas, such as the Poisson and Erlang B formulas used
for terrestrial links [21]. But with the increase of data services (e.g. corporate
networks interconnection, retail point-of-sale transactions, SCADA), it was nec-
essary to assign satellite resources in a more dynamic way, i.e. on a packet basis
rather than call (or circuit) basis. Demand assignment multiple access (DAMA)](https://image.slidesharecdn.com/resolutionforrandomaccessschemes-220518201640-39d36c53/85/Resolution-for-Random-Access-Schemes-pdf-20-320.jpg)
![1.1. MOTIVATION 7
schemes evolved to introduce faster capacity reservation mechanisms in order
to reduce end-to-end transmission times. In the CPODA protocol [24] capac-
ity reservations can be implemented via contention mini slots and piggybacking
them in the header of the scheduled packet transmissions. The CFDAMA-PB
protocol [25] behaves like the CPODA protocol, but in addition it randomly
assigns the unused traffic slots to inactive users. At low traffic loads the chance
that an earth station obtains free-assigned slots is high, thus reducing it’s end-
to-end transmission time to a minimum of one satellite hop propagation delay
(i.e. 250 ms for GEO satellites). At high loads, the end-to-end transmission de-
lay is not reduced as no spare satellite resources are left, but the system remains
stable
However, the performance of DAMA protocols is highly dependent on the
traffic characteristics and the number of earth stations sharing the satellite re-
sources. Internet data traffic is known to be statistically self-similar [22, 23].
Satellite terminals generate heavy tailed bursts of packets with a large variance
in the inter-arrival times between bursts. Besides, by aggregating streams of
such traffic from different earth stations we intensify the self-similarity (bursti-
ness) instead of smoothing it (fractal-like behavior). In [26] the performance of
the CFDAMA-PB protocol has been analyzed with traffic following a Poisson
regime, which is more appropriate for voice, facsimile or SCADA type of services,
but not for Internet data traffic as described above. In [27], the author studies
the sensitivity of the CFDAMA-PB protocol under different types of Internet
data traffic and a large number of earth stations (e.g. larger than the number of
traffic slots in one satellite hop propagation delay). It is shown that under bursty
traffic and a large number of earth stations the performance of CFDAMA-PB
is equivalent to a pure DAMA scheme, i.e. the majority of the free capacity
assignments are wasted. Figure 1.1 shows that the throughput component for
Free-Assigned slots with bursty traffic is very low at all loads, while for Poisson
traffic is quite high at loads up to 80%. Under these conditions, every packet
transmission undergoes first a reservation cycle (two satellite hops) and then
another satellite hop for the packet transmission (i.e. ≥ 750 ms for a GEO sce-
nario) like in a pure DAMA scheme. This poses some limitations on the use of
DAMA for consumer broadband access to support real-time or interactive type
of applications.
To enhance the performance of DAMA protocols for broadband satellite net-
works, predictive DAMA protocols have been developed, such as PRDAMA [28].
PRDAMA estimates the positive varying trend of the traffic on each earth sta-
tion to allocate the free capacity assignments. Non-linear prediction methods
are used to predict the traffic burstiness [29]. This results in a more efficient as-
signment of the free bandwidth to those terminals that are predicted to need the
resource, thus minimizing the wastage of free bandwidth assignments. However,
the effectiveness of the predictive methods cannot be guaranteed for all traffic
data [29]. Besides, the improvement in the end-to-end delay with respect to
CFDAMA-PB is rather limited (i.e. ∼ 100 ms delay reduction) with an average
end-to-end delay over a bent pipe GEO satellite in the order of 750 ms.
In [30] and [31] two combined random/reservation multiple access schemes](https://image.slidesharecdn.com/resolutionforrandomaccessschemes-220518201640-39d36c53/85/Resolution-for-Random-Access-Schemes-pdf-21-320.jpg)
![1.1. MOTIVATION 9
1.1.3 Key terrestrial random access techniques and their appli-
cability to satellite
RA protocols originated in the 1970s from the need for terminal-computer and
computer-computer communication. In computer communication, data traffic is
bursty. Traffic burstiness is the result of the high degree of randomness seen in
the message generation time and size, and of the relatively low-delay constraint
required by the user. Users generate traffic with a low duty cycle, but when they
do, they require a fast response. As a result, there is a large peak-to-average
ratio in the required data transmission rate. In this context, it is not efficient to
use fixed channel allocation schemes, as they would result in a low channel uti-
lization. A more advantageous approach is to provide a single shared high-speed
channel to the large number of users. However, when dealing with shared media
conflicts arise when more than one user want to access the shared resources si-
multaneously. Therefore, the challenge with shared media is to control the access
to the common channel while providing a good level of performance and main-
taining a reduced implementation complexity. RA protocols where developed to
address this multiple access scenario, and today are extensively used in terres-
trial networks over wired and wireless shared media [14, 32, 33]. In this section,
we review the key terrestrial RA techniques and analyze their applicability to
the satellite environment.
One of the most widely used distributed packet access schemes is the Car-
rier Sense Multiple Access (CSMA) and its variants. In CSMA, a station senses
the medium before transmitting and defers to any ongoing transmission [34].
CSMA/ Collision Detection (CSMA/CD) operates similarly to CSMA, but once
the transmission has started, if the sender detects a collision it stops transmitting
to reduce the overhead of a collision. When collisions occur, each station willing
to transmit backs off for a random time period [35]. Collision detection cannot
be implemented over a terrestrial wireless network due to the hidden terminal
problem where some stations are out of the transmission and detection range of
each other [36]. However it performs well in wired networks and the IEEE has
standardized CSMA/CD in the IEEE 802.3 standard [37]. Another variant used
in wireless networks is to avoid the collisions similarly to the CSMA/Collision
Avoidance (CSMA/CA) scheme. In this scheme the sender tries to avoid a colli-
sion after the channel becomes idle, by waiting for an Inter Frame Spacing (IFS)
time before contending for the channel. The IEEE has standardized CSMA/CA
in the IEEE 802.11 standard [38]. The back-off algorithm in CSMA/CA tries to
avoid collisions, but does not remove them all. Small values of the random back-
off time cause many collisions while very large values can cause unnecessarily
long delays. Secondly, CSMA/CA has failed to solve the hidden terminal prob-
lem, and cannot always detect that the medium is busy, thus creating a collision
in the channel. All the previous multiple access protocols employ carrier sensing
to avoid collisions and offer a good channel utilization, low latency and good
stability over channels where packet transmission times are larger than prop-
agation delays, but unfortunately cannot operate over satellite channels where
propagation delays are very large.](https://image.slidesharecdn.com/resolutionforrandomaccessschemes-220518201640-39d36c53/85/Resolution-for-Random-Access-Schemes-pdf-23-320.jpg)
![10 CHAPTER 1. MOTIVATION AND GOALS
Another type of schemes are distributed reservation schemes, also adopted in
terrestrial wireless networks. Reservations can be made either through a hand-
shaking on the same channel, or through an out-of-band signaling. There is a
large variety of hand-shaking protocols. The simplest one is the Multiple Ac-
cess Collision Avoidance (MACA) [39] protocol, in which a sender transmits a
Request to send (RTS) message to its intended receiver before the data trans-
mission. The data is transmitted only after reception of a Clear To Send (CTS)
message from the receiver, which the receiver sends on reception of a successful
RTS. The MACAW protocol [40] is an extension of the MACA protocol that in-
troduces enhancements to the RTS-CTS handshaking by adding a Data Sending
(DS) frame by the sender prior to data transmission and an Acknowledgement
frame (ACK) by the receiver following the data transmission with the aim of
solving the hidden terminal problem. The Floor Acquisition Multiple Access
(FAMA) [41] protocol combines carrier sensing with RTS-CTS handshaking,
again with the aim of solving the hidden terminal problem.
An example of out-of-band signaling protocol is the Busy Tone Multiple
Access (BTMA) [36]. In this scheme, a station that is within range and in line-
of-sight of all terminals, transmits a busy tone signal on a dedicated busy tone
channel, as long as it senses a carrier on the incoming message channel. It is by
sensing the busy tone channel that terminals determine the state of the message
channel. This scheme solves the hidden terminal problem, but introduces the
constraint that a control station needs to be within range of all terminals in the
network. Another type of out-band signaling distributed reservation scheme is
the Random Access Channel (RACH) used in the third generation (3G) cellular
networks [42], where terminals randomly transmit first short packet preambles,
and wait for a positive acquisition indicator from the base station prior to the
transmission of the complete message (i.e. after successful reservation of the
channel). Distributed reservation schemes reduce or eliminate the hidden ter-
minal problem typically present in terrestrial radio links, but also rely on short
propagation delays, i.e. the reservation delay only represents a small overhead
of the total packet transmission time. In the satellite environment, centralized
reservation schemes are used instead [25] to avoid several failed attempts prior
to the packet transmission, which could represent a very large overhead to the
total end-to-end packet transmission delay.
The pure ALOHA protocol, first proposed by N. Abramson in [43], is one of
the oldest and simplest multiple access protocols. In ALOHA, a terminal trans-
mits a packet without checking if any other terminal is active. Within an ap-
propriate timeout period, it receives and acknowledgment from the destination,
knowing that no conflict has occurred. Otherwise, it assumes that a collision has
occurred and must retransmit. To avoid continuously repeated collisions, the re-
transmission time is randomized across the terminals, thus spreading the retry
packets over time. A slotted version, referred to as Slotted Aloha (S-ALOHA) is
obtained by dividing time into slots of duration equal to the duration of a fixed-
length packet [44]. Users are required to synchronize the start of transmission of
their packets to the slot boundaries. When two packets collide they will overlap
completely rather than partially, providing an increase on channel efficiency over](https://image.slidesharecdn.com/resolutionforrandomaccessschemes-220518201640-39d36c53/85/Resolution-for-Random-Access-Schemes-pdf-24-320.jpg)
![1.1. MOTIVATION 11
pure Aloha. The S-ALOHA has the advantage of higher efficiency, but requires
time slot synchronization.
Both schemes are applicable to the satellite environment, as they have no
dependency on the propagation delay. Unfortunately, these schemes are subject
to a high collision probability (i.e. no carrier sensing), and their operation in
the high load region is not practical in the satellite environment due to the high
number of retransmissions required yielding very high latencies. The Diversity
Slotted Aloha (DSA) [46] is slightly improving the S-ALOHA performance at low
channel loads by sending twice the same packet at random locations in order to
increase the time diversity and thus reducing the Packet Loss Ratio (PLR). As
for S-ALOHA, operation in the high packet collision probability region is not
practical in a satellite environment.
It is generally assumed that whenever two packet transmissions overlap in
time, these packets destroy each other. This assumption is pessimistic as it
neglects capture or near-far effects in radio channels. Capture occurs when a
station receives messages simultaneously from two stations, but the signal from
one of them drowns out the other, so that no collision occurs. The station with
the higher received power is said to have captured the receiver. Some of these
effects have been addressed in [44, 45]. Capture is good in the sense that it
reduces the time needed in resolving collisions, but may also let weak terminals
completely out of the medium. However, in satellite networks the near-far effect
is typically very limited.
Figs. 1.2 and 1.3 present the performance results for S-ALOHA and DSA in
the presence of packets power imbalance following i.i.d. lognormal distributions
with equal mean µ and standard deviation σ with both parameters expressed
in dB in the logarithmic domain. The results have been obtained by detailed
simulations and by using the analytical model derived by the author in [47]. The
x-axis represents the normalized average channel MAC load (G) expressed in
information bits/symbol, in order to avoid any dependence with the modulation
cardinality or coding rate used. As we can see, in both schemes the throughput
improves with increasing power imbalance as collisions become easier to resolve
(power capture effect). However, as expected, the packet loss ratio is not low
and quickly increases as we increase the load on the channel. It can be remarked
that for low loads (e.g. G < 0.2), DSA outperforms S-ALOHA. This can be
better appreciated in Fig. 1.4, where S-ALOHA and DSA packet loss ratio
curves are combined in one figure for the case of no power imbalance and the
low load region is expanded. For instance, for a target PLR = 10−2, DSA can
achieve a throughput T = 0.05 while for S-ALOHA the maximum achievable
throughput is T = 0.01. This is justified by the fact that under light traffic
multiple transmission gives better PLR performance.
Slotted RA systems require terminals to keep the time slot synchronization.
The resulting synchronization overhead greatly reduces the system efficiency, in
particular for networks characterized by a large number of terminals with a very
low transmission duty cycle like the case in some of the envisaged applications
(M2M). Thus, slotted RA is penalizing low-cost terminal solutions. To mitigate
this limitation, a pure ALOHA scheme can be employed, but its performance](https://image.slidesharecdn.com/resolutionforrandomaccessschemes-220518201640-39d36c53/85/Resolution-for-Random-Access-Schemes-pdf-25-320.jpg)
![12 CHAPTER 1. MOTIVATION AND GOALS
0 0.5 1 1.5 2
0
0.1
0.2
0.3
0.4
0.5
0.6
Average MAC Channel Load [bits/symbol]
Throughput
[bits/symbol]
Sim. σ=3 dB
Sim. σ=2 dB
Sim. σ=1 dB
Sim. σ=0 dB
Ana. σ=3 dB
Ana. σ=2 dB
Ana. σ=1 dB
Ana. σ=0 dB
0 0.5 1 1.5 2
10
−2
10
−1
10
0
Average MAC Channel Load [bits/symbol]
Packet
Loss
Ratio
[PLR]
Sim. σ=3 dB
Sim. σ=2 dB
Sim. σ=1 dB
Sim. σ=0 dB
Ana. σ=3 dB
Ana. σ=2 dB
Ana. σ=1 dB
Ana. σ=0 dB
Figure 1.2: Analytical vs simulation S-ALOHA performance for QPSK modu-
lation, 3GPP FEC r = 1/2, packet block size 100 bits, Es/N0 = 7 dB in the
presence of lognormal packets power imbalance with mean µ = 0 dB, standard
deviation σ and Poisson traffic.](https://image.slidesharecdn.com/resolutionforrandomaccessschemes-220518201640-39d36c53/85/Resolution-for-Random-Access-Schemes-pdf-26-320.jpg)
![1.1. MOTIVATION 13
0 0.5 1 1.5 2
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0.45
Average MAC Channel Load [bits/symbol]
Throughput
[bits/symbol]
Sim. σ=3 dB
Sim. σ=2 dB
Sim. σ=1 dB
Sim. σ=0 dB
Ana. σ=3 dB
Ana. σ=2 dB
Ana. σ=1 dB
Ana. σ=0 dB
0 0.5 1 1.5 2
10
−2
10
−1
10
0
Average MAC Channel Load [bits/symbol]
Packet
Loss
Ratio
[PLR]
Sim. σ=3 dB
Sim. σ=2 dB
Sim. σ=1 dB
Sim. σ=0 dB
Ana. σ=3 dB
Ana. σ=2 dB
Ana. σ=1 dB
Ana. σ=0 dB
Figure 1.3: Analytical vs simulation DSA performance for QPSK modulation,
3GPP FEC r = 1/2, packet block size 100 bits, Es/N0 = 7 dB in the presence
of lognormal packets power imbalance with mean µ = 0 dB, standard deviation
σ and Poisson traffic.](https://image.slidesharecdn.com/resolutionforrandomaccessschemes-220518201640-39d36c53/85/Resolution-for-Random-Access-Schemes-pdf-27-320.jpg)
![14 CHAPTER 1. MOTIVATION AND GOALS
0 0.05 0.1 0.15 0.2
10
−4
10
−3
10
−2
10
−1
10
0
Average MAC Channel Load [bits/symbol]
Packet
Loss
Ratio
[PLR]
DSA Sim. σ=0 dB
SA Sim. σ=0 dB
DSA Ana. σ=0 dB
SA Ana. σ=0 dB
Figure 1.4: Analytical vs simulation S-ALOHA and DSA performance compar-
ison in the low load region for QPSK modulation, 3GPP FEC r = 1/2, packet
block size 100 bits, Es/N0 = 7 dB in the presence of no power imbalance and
Poisson traffic.](https://image.slidesharecdn.com/resolutionforrandomaccessschemes-220518201640-39d36c53/85/Resolution-for-Random-Access-Schemes-pdf-28-320.jpg)
![1.1. MOTIVATION 15
is worse than for S-ALOHA increasing by two the packet collision probabilities
[44].
Direct-sequence spread spectrum (DSSS) multiple access is the most common
form of CDMA whereby each user is assigned a particular code sequence which
is modulated on the carrier with the digital data modulated on top of that
[8]. Users can transmit asynchronously. Even when the same code is used,
data can be received [48, 49]. The Spread Spectrum ALOHA (SSA) proposed
in [48] has potentially attractive features as it provides a higher throughput
capability than S-ALOHA for the same PLR target under equal power multiple
access conditions when adopting powerful physical layer FEC (e.g. coding rates
≤ 1/2) and low order modulations (e.g. BPSK, QPSK). In [27] the author shows
through simplified analyses that SSA throughput is critically dependent on the
demodulator signal-to-noise plus interference (SNIR) threshold. Results reported
in [27] indicate that differently from S-ALOHA, SSA shows a steep PLR increase
with MAC load. Thus SSA can be operated with low PLR close to the peak of
the throughput characteristic. As an example, using turbo codes and relatively
small packets, SSA can achieve throughput in the order of T = 0.5 bits/chip for
a packet loss ratio of 10−3 (see Fig. 1.5 with σ = 0 dB).
This is a very interesting random access scheme for the satellite environment,
in particular when asynchronous access is considered. The main reason for the
performance improvement of SSA techniques with regards to pure ALOHA is
that they can take advantage of a higher traffic aggregation. The average number
of packet arrivals over one packet duration can be computed as follows:
λ = Nrep G Gp, (1.1)
where Nrep is the number of replicas transmitted for each packet, G is the MAC
load expressed in information bits/symbol in non-spread systems and bits/chip
in spread systems. The processing gain is given by Gp = SF/(r log2 M) where
r is the channel coding rate, M is the modulation cardinality and SF is the
spreading factor.
It can be observed from eqn. (1.1) that large processing gains will increase
the value of λ. Typical values for non-spread spectrum systems are λ ≤ 5 while
for spread-spectrum systems λ ≈ 100. In Fig. 1.6 we can see that the Poisson
PDF normalized to the mean, approaches a delta as we increase λ. This means
that the instantaneous number of interfering packets will fluctuate less and less
as we increase λ. This is favorable in RA because we can easily make the system
to work around a desired load point (i.e. avoid having peaks of traffic) and also
because interference can be accurately approximated as AWGN when we have a
large number of interferer packets.
However, the SSA Achilles’ heel resides in its high sensitivity to multiple
access carrier power unbalance. This phenomenon is disrupting the SSA scheme
throughput. The SSA throughput is diminished by several orders of magnitude
when the received packets power is lognormally distributed with standard devi-
ation of 2-3 dB (see Fig. 1.5), as opposed to S-ALOHA and DSA that improved
performance with power imbalance due to the power capture effect. SSA is de-
signed to work with MAI. Optimal power distribution for SSA is achieved when](https://image.slidesharecdn.com/resolutionforrandomaccessschemes-220518201640-39d36c53/85/Resolution-for-Random-Access-Schemes-pdf-29-320.jpg)
![16 CHAPTER 1. MOTIVATION AND GOALS
all packets have the same power level [27, 50]. When there is power unbalance
the weakest packets may become not decodable (unless some interference can-
cellation technique is applied), thus reducing the RA throughput. DSA and
S-ALOHA are designed to work with no or very little interference (collisions
are in principle destructive). Under power imbalance some collisions can be
recovered (capture effect), thus slightly improving the performance.
The above review of known terrestrial RA techniques reveals that none of
them is fully performing over the satellite environment. Table 1.1 provides a
summary of the different terrestrial RA techniques analyzed in this section. In
general, terrestrial RA schemes are known to provide low channel utilization over
the satellite environment due to the long propagation delays. Among all of them,
the ALOHA-based techniques adapt better to the satellite environment, as they
do not have any dependency on the propagation delay, but their performance is
constrained by their high collision probability. As a result, today’s satellite sys-
tems only use ALOHA-based RA for initial network login, capacity request and
short packet transmissions [51, 52]. Therefore, it can be concluded that there
is a need to develop new high-performance RA protocols to satisfy the grow-
ing market demands for satellite broadband access and M2M communications
described in Sects. 1.1.1 and 1.1.2.
Technique Main Characteristics
Carrier Sense Multiple Access (CSMA) Carrier sense, reduced collision probability,
sensitivity to propagation delay.
CSMA Collision Detection (CSMA/CD) CSMA with reduced collision overhead.
CSMA Collision Avoidance (CSMA/CA) CSMA with collision avoidance mechanism.
Multiple Access Collision Avoidance (MACA) Distributed reservation, sensitivity to propagation delay.
MACA for Wireless (MACAW) Distributed reservation, sensitivity to propagation delay.
Floor Acquisition Multiple Access (FAMA) Distributed reservation and carrier sense, sensitivity to
propagation delay.
Busy Tone Multiple Access (BTMA) Distributed reservation, sensitivity to propagation delay.
3G Random Access Channel (RACH) Distributed reservation, sensitivity to propagation delay.
Slotted ALOHA (S-ALOHA) High collision probability, no dependency propagation delay.
Diversity Slotted ALOHA (DSA) S-ALOHA with multiple copies, improved PLR under light
traffic.
Spread Spectrum ALOHA (SSA) Asynchronous CDMA, no dependency propagation delay,
improved PLR under equal packets power.
Table 1.1: Summary of Terrestrial Random Access Techniques.
1.2 Goals
As described in Sect. 1.1, the improvement of Random Access (RA) protocols
performance over satellite could bring great performance advantages to satellite
communication networks in serving the growing demand for satellite broadband
access and M2M communications. RA protocols provide fast access and are
insensitive to the network population size and the traffic characteristics. Unfor-
tunately, they offer poor channel utilization over the satellite environment due
to the long propagation delays, as well as low reliability due to the high collision
probabilities. As we have seen in Sect. 1.1.3, there is a big margin for improve-
ment of RA schemes over satellite networks in terms of bandwidth utilization,
which is typically not exceeding 0.05 bit/s/Hz if we want to achieve a target
Packet Loss Ratio (PLR) < 10−2.](https://image.slidesharecdn.com/resolutionforrandomaccessschemes-220518201640-39d36c53/85/Resolution-for-Random-Access-Schemes-pdf-30-320.jpg)
![1.2. GOALS 17
0 0.2 0.4 0.6 0.8 1 1.2 1.4
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Average MAC load (bits/chip)
Throughput
[bits/chip] σ=0dB, semi−ana
σ=1dB, semi−ana
σ=2dB, semi−ana
σ=3dB, semi−ana
σ=0dB, simulated
σ=1dB, simulated
σ=2dB, simulated
σ=3dB, simulated
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
10
−8
10
−7
10
−6
10
−5
10
−4
10
−3
10
−2
10
−1
10
0
Average MAC load (bits/chip)
Packet
Loss
Ratio
(PLR)
σ=0dB, semi−ana
σ=1dB, semi−ana
σ=2dB, semi−ana
σ=3dB, semi−ana
σ=0dB, simulated
σ=1dB, simulated
σ=2dB, simulated
σ=3dB, simulated
Figure 1.5: Simulated vs. Analytical SSA performance with and without power
unbalance from [50]: 3GPP FEC r = 1/3 with block size 100 bits, BPSK mod-
ulation, spreading factor 256.](https://image.slidesharecdn.com/resolutionforrandomaccessschemes-220518201640-39d36c53/85/Resolution-for-Random-Access-Schemes-pdf-31-320.jpg)
![Chapter 2
The work
In this thesis, the author presents new advanced random access techniques de-
rived from the ALOHA protocols that boost their performance over the satellite
environment and enable new market opportunities for satellite communications.
As we have seen in Sect. 1.1.3, the RA techniques that best adapt to the satellite
environment are the ALOHA-based techniques, as they do not have any depen-
dence on propagation delay. This imposes that the new RA schemes shall be
able to operate under very high packet collision conditions.
To achieve that goal we shall exploit the advances in signal processing tech-
niques such as interference cancellation (IC) and strong forward error correction
(FEC) coding techniques. The multiuser detection research area covers many
different algorithms for mitigating interference between users [53]. While clas-
sical IC often assumes synchronous CDMA continuous transmission, this thesis
addresses the cases of IC for asynchronous bursty transmissions (spread and
non-spread spectrum based), as well as TDMA-based systems. This thesis in-
vestigates coordinated IC processing techniques at a common receiver (e.g the
hub) with the aim of achieving an aggregated data rate of all simultaneous users
approaching the Shannon capacity of the Gaussian noise channel, even for the
asynchronous case. The choice of the physical layer FEC scheme for RA systems
is not as straight-forward as for conventional communication systems operating
in Additive White Gaussian Noise (AWGN). Being a RA Multiple Access (MA)
system, the individual but also the aggregate MA performance are relevant. FEC
coding schemes shall operate in front of strong interference and not always under
AWGN-like conditions (e.g. for TDMA based systems). Therefore lower coding
rates, although apparently reducing the individual packet bit rate, may be im-
portant to enhance the overall RA scheme performance. This thesis investigates
the impact of the FEC coding rate on the RA protocols performance. As ex-
plained in Sect. 1.1.3, the level of traffic aggregation in TDMA-based systems is
much lower than in CDMA ones. This means that the instantaneous number of
interfering packets in TDMA-based systems fluctuates a lot and we cannot guar-
antee to operate under a desired working point. This thesis investigates novel
approaches to introduce time diversity in TDMA-based systems in order to im-
prove the collision probability conditions. The complexity of the new proposed
techniques is analyzed to ensure they are commercially viable.
19](https://image.slidesharecdn.com/resolutionforrandomaccessschemes-220518201640-39d36c53/85/Resolution-for-Random-Access-Schemes-pdf-33-320.jpg)
![20 CHAPTER 2. THE WORK
This thesis also analyses the performance of the new proposed RA tech-
niques and compares them with traditional RA protocols. RA analytical results
for conventional RA schemes presented so far were assuming uncoded systems
and equi-powered interfering packets [43, 44], which is not what is required in
modern satellite communication systems. Some analysis of the power capture
effect is reported in [45, 54, 55, 56], but existing models are limited to discrete
power levels, do not provide an accurate modeling of the RA Forward Error Cor-
rection (FEC) behavior, and more importantly, do not include IC. This thesis
defines new RA analytical frameworks that allow calculating the performance
of traditional and new RA techniques in terms of PLR and throughput. The
proposed frameworks assume arbitrary power and traffic distributions, and ac-
curately model the RA FEC and IC behavior. Analytical results have been
successfully compared to a wide set of simulation results showing the capability
of the proposed analytical framework to predict the the RA scheme performance
in the vast majority of application cases.
The contributions have been grouped into slotted and asynchronous RA tech-
niques and are summarized in the following sections.
2.1 Slotted Random Access
Research for slotted RA techniques took place between 2005 and 2009, and then
complemented by an analytical framework in 2014. The research on slotted sys-
tems was triggered by the need to introduce efficient and reliable RA schemes to
the existing satellite systems using standards for interactive satellite broadband
networks like the Digital Video Broadcasting (DVB) Return Channel via Satel-
lite (DVB-RCS) [51] and the Telecommunication Industry Association (TIA)
IP over Satellite (IPoS) [52]. The standard-based and many proprietary satellite
communication systems employ MF-TDMA access and rely primarily on DAMA
schemes for the transmission of data traffic on the return link (earth station to
hub), as described in Sect. 1.1.2. The arrival of broadband Internet access for
the consumer market introduced the need for a fast satellite access to a very
large population of earth stations sharing the same bandwidth.
The first contribution from the author is a RA scheme dubbed Contention
Resolution Diversity Slotted Aloha (CRDSA) [57] and included in this thesis in
chapter 4. The CRDSA technique consists in transmitting every burst two or
more times (Nrep) in randomly selected slots within a TDMA frame of Nslots.
Each burst contains in the header information of the location of the replicas
within the frame (see Fig. 2.1). The transmitter achieves network synchro-
nization in the same way as it is defined in the MF-TDMA based standards
[51, 52], i.e. the terminal exploits the forward link network clock reference re-
quired for MF-TDMA terminal time reference generation to also generate time
and frequency reference for CRDSA.
The complete TDMA frame is sampled and stored in a digital memory at
the receiver side. By using simple yet efficient interference cancellation tech-
niques, clean bursts are recovered (e.g. packet 3 in slot 5 in Fig. 2.1) and the
interference generated by their replicas on other slots are cancelled (e.g. packet](https://image.slidesharecdn.com/resolutionforrandomaccessschemes-220518201640-39d36c53/85/Resolution-for-Random-Access-Schemes-pdf-34-320.jpg)
![2.1. SLOTTED RANDOM ACCESS 21
2 2
1 3
3 4 4 6
5
6
5 1
TDMA RA frame with Nslots
TDMA RA frame with Nslots
RA slot
RA slot
Figure 2.1: Example of CRDSA frame with two replicas per packet.
3 in slot 4). By performing an iterative processing of the TDMA frame it is
proven that most of the initial collisions can be resolved. In Fig. 2.2 it is shown
that CRDSA largely outperforms classical Slotted Aloha techniques in terms of
throughput and packet loss ratio (PLR). For a MAC packet loss probability of
2%, channel utilization efficiencies close to 40% can be achieved with the CRDSA
technique (see Fig. 2.2(b)), while for S-ALOHA only a channel utilization effi-
ciency of 2% can be achieved for the same packet loss probability (see Fig. 1.4).
This represents a 20-fold throughput improvement compared to S-ALOHA. It
is important to highlight the importance of PLR improvement in the satellite
environment, as it has a direct benefit on the end-to-end delay due to a reduction
in the need for packet retransmissions. The author’s contribution [57] presents
detailed performance results on delays.
The performance boost of the CRDSA protocol is achieved thanks to the im-
plementation on the hub receiver of a collision resolution capability that exploits
IC techniques. IC techniques have been largely investigated for Code Division
Multiple Access (CDMA) [53] but, at the author’s knowledge, have never been
proposed in a TDMA S-ALOHA context. CRDSA is basically a TDMA access
scheme that uses the information from the successfully decoded packets to cancel
the interference their replicas may generate on other slots. We are able to cancel
the interference because we know the data symbols of the interference from the
successfully decoded packets. One of the main issues in applying IC techniques
to TDMA S-ALOHA is related to the need for accurate channel estimation for
the burst replicas removal where collision(s) occur. In fact, collisions in a satel-
lite TDMA multiple access channel are typically destructive as the near-far effect
is rather limited. While carrier frequency, amplitude and timing estimation for
IC can be derived from the ”clean” replica (e.g. packet 3 in slot 5 in the example
from Fig. 2.1), carrier phase has to be estimated on the slot where collision(s)
occurs (e.g. in slot 4). This is because the phase in practical broadband systems
is time variant also from slot to slot. This key problem has been solved in [57]
by exploiting the burst preamble which is individually ”signed” by a pseudo-
random binary sequence randomly selected among the available code family by
each active earth station for each burst in each frame. All replicas of the same
burst use the same preamble code. This approach does not require a centralized
preamble code assignment, thus allowing to maintain the random access nature
of the proposed scheme. In this way the preamble can be used for carrier phase](https://image.slidesharecdn.com/resolutionforrandomaccessschemes-220518201640-39d36c53/85/Resolution-for-Random-Access-Schemes-pdf-35-320.jpg)
![2.1. SLOTTED RANDOM ACCESS 23
estimation also in case of multiple collisions which normally are destructive for
channel estimation and payload decoding. With the above channel estimation
implementation, it is shown in Section V.C in from [57] that CRDSA perfor-
mance results on non-ideal channel estimation for IC are very close to those
derived with ideal channel estimation.
An alternative way to perform channel estimation is to use a single Unique
Word (UW) for all bursts and all earth stations, but use some of the payload
symbols, which are known from the clean replicas, for carrier phase estimation.
Payload symbols are uncorrelated among the colliding packets, even though with-
out excellent cross-correlation properties. It is found that by using payload sym-
bols for carrier phase estimation, performance results are comparable to those
obtained by using multiple pseudo-random binary sequences for the burst pream-
bles. But his approach reduces complexity on both the earth stations and hub
as a single UW is used. Besides, the knowledge of the payload data allows to
track the carrier phase along the burst by using an normal Data Aided-Phase
Locked Loop (DA-PLL), which is particularly useful when low cost terminals
with very low baud rate are used (e.g for interactive TV), since the RF front
ends introduce a phase noise (nuisance) making the carrier phase not constant
over the burst duration. This improvement was proposed by Daniel Delaru-
elle from NEWTEC and Jean Pierre Choffray from SES-ASTRA. Both channel
estimation approaches have been extensively investigated in [58] (see chapter 6).
Prior to the IC of a packet, the whole packet is re-modulated allowing data
aided refined amplitude and phase estimation. This maximizes the IC efficiency
with a minimal degradation of the CRDSA performance when compared to ideal
channel estimation conditions [58].
CRDSA has a number of parameters that can greatly influence their per-
formance, being the most important ones FEC coding rate ρ, the number of
replicas Nrep and the packets power imbalance. With the aim of optimizing
its performance, the author has developed an analytical framework and a de-
tailed simulator for the performance assessment of slotted RA protocols (see
[47] included in this thesis in chapter 5). The proposed analytical framework
assumes arbitrary power and traffic distributions, and accurately models the
FEC and RA IC behavior. To be remarked that using a powerful FEC code
(e.g. ρ = 1/3) and an adequate signal to noise ratio (e.g. SNR = 10 dB), there
is non zero probability of correctly detecting the packet even in presence of a
collision (capture effects). Besides FEC coding schemes shall operate in front of
strong interference and not always under AWGN-like conditions (e.g. for TDMA
based systems). When payload burst are protected by a powerful FEC, a phase
estimation error standard deviation up to 15 degrees can be tolerated without
significant impact on the performance [58]. Therefore lower coding rates, al-
though apparently reducing the individual packet bit rate, are very important
to enhance the overall RA scheme performance. Three burst replicas gives bet-
ter performance than two, because it significantly reduces the loop occurrence
probability between packets from different earth stations. In the example if Fig.
2.1, packets 4 and 5 have formed a loop as their replicas have been transmitted
in the same slots. The issue of loops has been extensively analyzed by the author](https://image.slidesharecdn.com/resolutionforrandomaccessschemes-220518201640-39d36c53/85/Resolution-for-Random-Access-Schemes-pdf-37-320.jpg)
![24 CHAPTER 2. THE WORK
in [47], and with three packet replicas the probability of occurrence is well below
the packet loss ratio target for our applications (PLR < 10−3). Packets power
imbalance greatly contributes to the interference cancellation process and boosts
the performance of CRDSA. Strong packets are decoded first and weaker packets
are successively decoded following the iterative IC process. Power imbalance re-
sults from the combined effect of the time variant atmospheric propagation, open
loop power control errors (if applicable), earth station equivalent isotropically
radiated power (EIRP) and satellite receive antenna gain variations. It is shown
in Fig. 2.3 that CRDSA with FEC ρ = 1/3, Nrep = 3 and power imbalance
following a lognormal power distribution with standard deviation σ = 3 dB, can
achieve a throughput T = 1.4 bits/symbol for a PLR < 10−3, which represents
a 1400-fold improvement with respect to S-ALOHA.
2.2 Asynchronous Random Access
Research on asynchronous random access took place between 2008 and 2014.
The research on asynchronous access was triggered by the growing demand for
low cost and highly efficient solutions supporting both fixed and mobile satellite
services, such as consumer broadband access, machine-to-machine communica-
tions (M2M), interactive TV, supervisory control and data acquisition (SCADA),
transaction and safety of life applications. These networks are generally char-
acterized by a large population of low power and small size terminals, sharing
the available resources under very dynamic traffic conditions. In particular, in
the return link (earth station to hub) users are likely to generate low duty cy-
cle bursty traffic with extended inactivity periods. Slotted RA systems require
terminals to keep the time slot synchronization through a control loop. The
resulting synchronization overhead can greatly reduce the system efficiency, in
particular for networks characterized by a very large number of terminals with a
very low transmission duty cycle, like it is the case in some of the envisaged appli-
cations (e.g. M2M). Thus, slotted RA is penalizing low-cost terminal solutions.
To mitigate this limitation, asynchronous RA schemes are preferable.
Asynchronous RA schemes can been divided into spread spectrum and non-
spread spectrum multiple access techniques. This thesis presents two novel
asynchronous RA schemes. The first one, dubbed Enhanced Spread Spectrum
ALOHA (E-SSA), is an extension of Spread Spectrum ALOHA presented in
Section 1.1.3 and introduces an IC mechanism at the receiver. The second RA
technique, Asynchronous Contention Resolution Diversity ALOHA (ACRDA),
is an extension of the CRDSA protocol presented in Sect. 2.1 to a truly asyn-
chronous multiple access environment. ACRDA is non-spread spectrum based,
but incorporates some of the E-SSA receiver processing functions to support the
asynchronous access from all the earth stations to the shared medium.
2.2.1 Enhanced Spread Spectrum ALOHA
Enhanced Spread Spectrum ALOHA (E-SSA) behaves like SSA on the trans-
mitter side (see Sect. 1.1.3). The E-SSA novelty lies on the detector [59]. The](https://image.slidesharecdn.com/resolutionforrandomaccessschemes-220518201640-39d36c53/85/Resolution-for-Random-Access-Schemes-pdf-38-320.jpg)
![2.2. ASYNCHRONOUS RANDOM ACCESS 25
0 0.5 1 1.5 2
0
0.5
1
1.5
Average MAC Channel Load − G [b/s/Hz]
Normalized
Throughput
[T] Sim. σ=3 dB
Sim. σ=2 dB
Sim. σ=1 dB
Sim. σ=0 dB
Ana. σ=3 dB
Ana. σ=2 dB
Ana. σ=1 dB
Ana. σ=0 dB
(a) Throughput
0 0.5 1 1.5 2
10
−7
10
−6
10
−5
10
−4
10
−3
10
−2
10
−1
10
0
Average MAC Channel Load − G [b/s/Hz]
Packet
Loss
Ratio
[PLR]
Sim. σ=3 dB
Sim. σ=2 dB
Sim. σ=1 dB
Sim. σ=0 dB
Ana. σ=3 dB
Ana. σ=2 dB
Ana. σ=1 dB
Ana. σ=0 dB
(b) PLR
Figure 2.3: Analytical vs. Simulation CRDSA performance for Nrep = 3, Niter
max =
15, Nslots=1000, QPSK modulation, 3GPP FEC ρ = 1/3, packet block size 100
bits, Es/N0 = 10 dB in the presence of lognormal packets power unbalance with
mean µ = 0 dB, standard deviation σ and Poisson traffic [47].](https://image.slidesharecdn.com/resolutionforrandomaccessschemes-220518201640-39d36c53/85/Resolution-for-Random-Access-Schemes-pdf-39-320.jpg)
![26 CHAPTER 2. THE WORK
4000 4050 4100 4150 4200 4250 4300
0
100
200
300
400
500
600
Time
Number
of
interfering
packets
Packet duration=312; Normalized Load=0.5
Figure 2.4: Realization of the instantaneous and average number of interfering
packets over packet of interest [60].
detector located at the hub is the heart of the system, as it has to support a
high throughput of incoming packets. For a typical spread spectrum system with
Nrep = 1, SF = 256, r = 1/3, M = 2 and a normalized load G = 0.5, the aver-
age number of packet arrivals over a packet duration is λ = 384 (see eqn. (1.1)),
which represents quite a large value. In Fig. 2.4, the author shows that the
instantaneous number of interfering packets is almost constant around the mean
value λ = 384 (red line) over the whole packet duration. This is advantageous
for a RA scheme design because the interference can be accurately approximated
as AWGN when we have a large number of interferer packets. The time on the
x-axis corresponds to symbol time steps, and the packet shown in this example
is a 100 bit packet with 300 coded bits. The complete set of system parameters
used for the E-SSA analyses can be found in Table I from [50], also included in
this thesis in chapter 7.
The principle of the E-SSA detector is illustrated in Fig. 2.5. The received
CDMA signal is band-pass filtered, sampled, digitally down-converted to base-
band with I-Q components separation and stored in a digital memory of 2·WNc
s
real samples. We assume that complex baseband samples are stored in I-Q for-
mat. Nc
s corresponds to the number of chips per physical layer channel symbol
and W corresponds to the memory window size in symbols for the interference
cancellation process at the hub. The window size W shall be optimized to be
the smallest possible value yielding good IC performance. Typically, W should](https://image.slidesharecdn.com/resolutionforrandomaccessschemes-220518201640-39d36c53/85/Resolution-for-Random-Access-Schemes-pdf-40-320.jpg)
![28 CHAPTER 2. THE WORK
A detailed analytical description of the E-SSA algorithm is provided in Ap-
pendix I from [50], also included in chapter 7 of this thesis. Two important steps
for an optimal E-SSA performance are the packet preamble detection in front
of very high interference (step 2) and the IC process defined in step 5(c). The
hub demodulator starts searching for the presence packet preamble by means
of a conventional preamble correlator. Because of the incoming packets car-
rier frequency uncertainty due to the oscillator instabilities, preamble parallel
search in the frequency domain is typically required. The preamble acquisition
performance needs to be lower than the target PLR (e.g 10−3) in terms of the
probability of missed detection Pmd and the probability of false alarm Pfa at the
very high MAC operating load allowed by E-SSA. Preamble acquisition aspects
have been summarized in Sect. IV.E in [50]. A particular aspect in SS is the
need for a longer preamble for packet detection due to operation under very
high MAI (Ec/(N0 + I0) ≈ −30 dB). When a preamble is detected, the associ-
ated packet amplitude, carrier frequency, chip timing and phase are estimated
exploiting the preamble plus possible auxiliary pilot symbols. At this point, de-
spreading, coherent detection, payload and signalling bits decoding takes place.
This demodulation step is comparable to a conventional DS-SS burst demodu-
lator. The transmitted packet contains also a CRC, which is exploited by the
demodulator to verify that the packet has been correctly detected. If this is the
case, the payload information is extracted. At the same time the decoded bits are
re-encoded and re-modulated. This locally regenerated packet is then correlated
with the memory samples corresponding to the currently detected packet loca-
tion. Through this correlation process it is possible to extract a more accurate
amplitude and carrier phase estimation compared to the initial demodulation
step. This is because the replica packet(s) amplitude and phase estimation is
now extending over the whole packet duration instead of being limited to the sole
preamble length. In this way, amplitude and phase variations over the packet
duration can be estimated resulting in a more accurate cancellation process.
This key step is called payload data-aided refined channel estimation and is the
same used for CRDSA IC [58]. Thanks to the refined channel estimation, the
E-SSA demodulator achieves accurate cancellation of the detected packet from
the sliding window memory.
The E-SSA performance has been investigated in-depth both by analysis and
simulation in Sect. IV.D from [50]. The analytical framework proposed by the
author, accurately models the performance of spread-spectrum RA techniques,
such as SSA and E-SSA, and assumes arbitrary power and traffic distributions,
and accurately models the FEC and RA IC behavior (see Section IV.A from [50]).
Fig. 2.6 reports some key conclusion about E- SSA performance. Performance
results shall be compared to the SSA ones reported in Fig. 1.5. First of all,
assuming a target PLR of 10−3 and no power imbalance, the E-SSA throughput
is 1.12 bits/symbol, i.e. 2.4 times higher than conventional SSA. When lognor-
mal distributed packet power is assumed (with standard deviation σ = 2 dB),
then the E-SSA throughput reaches 1.7 bits/symbol, i.e. 17 times larger than
SSA. This striking result is due to the iSIC superior performance compared to a
conventional SSA burst demodulator in case of imbalanced power. The received](https://image.slidesharecdn.com/resolutionforrandomaccessschemes-220518201640-39d36c53/85/Resolution-for-Random-Access-Schemes-pdf-42-320.jpg)
![2.2. ASYNCHRONOUS RANDOM ACCESS 29
packets power fluctuation at the hub is caused by the uplink Land Mobile Satel-
lite (LMS) channel shadowing and multipath fading, as well as the satellite and
earth station antenna gain variations as a function of the location. In Sect. VI
from [50], Riccardo De Gaudenzi proposes a SNIR driven uplink packet transmis-
sion control (SDUPTC). SDUPTC is based on the principle to transmit packets
only when the downlink signal quality is good enough, i.e. the signal strength
or better SNR is within a certain window representative of Line-of-Sight (LOS)
conditions. If this is not the case the transmission is delayed until (quasi)-LOS
conditions are verified. The proposed simple SDUPTC algorithm is intended
to maximize the probability of packet successful transmission, limit the power
imbalance among the packets received at the hub side to improve the RA scheme
performance and to avoid creating multiple access interference (MAI) when not
useful for packet transmission.
2.2.2 Asynchronous Contention Resolution Diversity ALOHA
The Asynchronous Contention Resolution Diversity ALOHA (ACRDA) proto-
col is analyzed in detail in [61], and included in this thesis in chapter 9. An
introduction to the ACRDA patent [62] is also presented in chapter 10. Unlike
S-ALOHA, DSA or CRDSA, ACRDA eliminates the need to maintain accurate
slot synchronization among all transmitters, but differently to SSA and E-SSA,
does not require the use of spread spectrum techniques. The need for transmit-
ter synchronization is a major drawback for very large networks (e.g. M2M), as
the signaling overhead scales up with the number of transmitters independently
from their traffic activity factor.
In ACRDA, earth stations behave like in CRDSA, exploiting packet replicas
and the associated location signaling, but do not need to maintain slot synchro-
nization with the hub. The term Virtual Frame (VF) is introduced here to refer
to the concept of frame of slots that is only local to each transmitter. In ACRDA,
for all transmitters, each VF is composed of a number of slots Nslots and each
slot has a duration Tslot with an overall frame duration Tframe = Nslots · Tslot.
Fig. 2.7 shows the Virtual Frame compositions for the ACRDA scheme. The
different transmitters are not time synchronized, and hence, the time offset be-
tween VF(i), VF(i − 1), VF(i + 1) is arbitrary. In case of mobile applications,
the Doppler effect may have an appreciable impact in terms of incoming packets
clock frequency offset. In this situation the VFs will have slightly different dura-
tion. However, the localization process of the replica packets within each VF will
remain accurate, as the hub burst demodulator will extract for each VF its own
clock reference. Since the demodulator has no knowledge of the start of VFs,
the signaling of the replicas location within the VF has to be relative to each
packet position and not absolute to the start of the VF. Two options exist for
the ACRDA modulator (as described in Sect. II.A from [61]), which consist in
transmitting all packet replicas randomized within the VF, or transmitting the
first packet replica in the first slot of the VF and the remaining replicas random-
ized within the VF. The latter option, improves end-to-end delay performance
since first packet replica is transmitted without any added delay.](https://image.slidesharecdn.com/resolutionforrandomaccessschemes-220518201640-39d36c53/85/Resolution-for-Random-Access-Schemes-pdf-43-320.jpg)
![30 CHAPTER 2. THE WORK
0 0.5 1 1.5 2 2.5
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
MAC load (bit/s/Hz)
Throughput
(bit/s/Hz) σ=0 dB, ana.
σ=1 dB, ana.
σ=2 dB, ana.
σ=3 dB, ana.
σ=0 dB, sim.
σ=1 dB, sim.
σ=2 dB, sim.
σ=3 dB, sim.
(a) E-SSA Throughput with and without power unbalance.
1 1.5 2 2.5
10
−8
10
−7
10
−6
10
−5
10
−4
10
−3
10
−2
10
−1
10
0
MAC load (bit/s/Hz)
Packet
Loss
Ratio
(PLR)
σ=0 dB, ana.
σ=1 dB, ana.
σ=2 dB, ana.
σ=3 dB, ana.
σ=0 dB, sim.
σ=1 dB, sim.
σ=2 dB, sim.
σ=3 dB, sim.
(b) E-SSA PLR with and without power unbalance.
Figure 2.6: Simulated vs. Analytical E-SSA performance with and without power
unbalance, 3GPP FEC r = 1/3 with block size 100 bits, BPSK modulation,
spreading factor 256 [50].](https://image.slidesharecdn.com/resolutionforrandomaccessschemes-220518201640-39d36c53/85/Resolution-for-Random-Access-Schemes-pdf-44-320.jpg)
![2.2. ASYNCHRONOUS RANDOM ACCESS 31
4 4 4
Virtual Frame (i-1)
5 5 5
Virtual Frame (i)
3 3 3
Virtual Frame (i-2)
6 6 6
Virtual Frame (i+1)
7 7 7
Virtual Frame (i+2)
2 2 2
Virtual Frame (i-3)
1 1
1
Virtual Frame (i-4)
8
Virtual Frame (i+3)
9
Virtual Frame (i+4)
9 9
8 8
∆W
∆W
W(s)
W(s)
W(s + 1)
W(s + 1)
Offset between
VF(i-1) and VF(i)
Offset between
VF(i) and VF(i+1)
Figure 2.7: Virtual frame compositions in ACRDA.
ACRDA combines features of CRDSA, and E-SSA on the receiver side. On
one hand, the same E-SSA sliding window-based memory processing described
in Sect. 2.2.1 is adopted to handle the asynchronously arriving packet replicas.
On the other hand, for a given receiver memory position, the replica packets
cancellation scheme described in Sect. 2.1 is borrowed from the CRDSA de-
modulator processing. The received signal is sampled at baseband, and complex
signal samples are stored in a sliding window memory of W virtual frames (see
Fig. 2.7). Typically W = 3 for optimal ACRDA performance, as for E-SSA.
For a given window position, the demodulator performs the same iterative pro-
cessing as for CRDSA to decode the clean packets and perform IC of the replica
packets. The detailed ACRDA demodulator operation is described in Sect. II.B
from [61]. Once the iterative processing of a specific window position is com-
pleted, the sliding window is shifted towards the right in time by ∆W, which is a
fraction of the VF duration (e.g. ∆W = 1/6). The oldest samples spanning the
leftmost part of the memory will be removed and the emptied rightmost part of
the memory will be then filled with the new incoming complex samples.
ACRDA provides better PLR and throughput performance than CRDSA.
Fig. 2.8 compares the performance of the ACRDA vs. CRDSA, evaluated via
mathematical analysis and computer simulations. Similarly to the other pro-
posed RA schemes, the author proposes in Sect. III from [61] an analytical
framework derived from the RA analytical framework presented in [47]. For
a target PLR = 10−3, ACRDA channel can be loaded up to 1 bits/symbol,
while CRDSA not more than 0.75 bits/symbol. It is important to note that the
PLR floor present at low loads (i.e. G < 1 bits/symbol) is lower for ACRDA
than CRDSA. The reason for this is that the probability of loops occurrence,](https://image.slidesharecdn.com/resolutionforrandomaccessschemes-220518201640-39d36c53/85/Resolution-for-Random-Access-Schemes-pdf-45-320.jpg)
![32 CHAPTER 2. THE WORK
described in Sect. 2.1, is significantly lower in ACRDA than in CRDSA due to
the asynchronous nature of the access. The effect of loops in performance is an-
alyzed in detail in Sect. III from [61] for ACRDA and in Appendix D from [47]
for CRDSA. To mitigate this limitation, the optimal configuration for CRDSA is
with Nrep = 3 as presented in Fig. 2.3. In ACRDA, slightly better performance
can be already achieved with Nrep = 2, thus reducing the demodulator com-
plexity. It shall be recalled that ACRDA also reduces earth station complexity,
as it does not require any control loop to keep time slot synchronization with
the hub. Similarly to CRDSA, ACRDA performance improves in front of power
imbalance. Fig. 8 from [61] shows that ACRDA with FEC ρ = 1/3, Nrep = 2
and power imbalance following a lognormal power distribution with standard
deviation σ = 3 dB, can achieve a throughput T = 1.5 bits/symbol for a target
PLR < 10−3.
In Sect. IV from [61], Guray Acar and Eloi Garrido analyze and compare the
delay performance between ACRDA and CRDSA. Results show that ACRDA
reduces the delay by a factor of 10 compared to CRDSA for low loads (e.g. G =
0.3 bits/symbol) and a factor of 2 at high loads (e.g. G = 0.9 bits/symbol).
In Fig. 2.9, the performance of ACRDA is compared against E-SSA. As we
can see, in particular for the case of large power imbalance, E-SSA performance
is better than ACRDA. The use of QPSK in ACRDA improves the efficiency,
but also increases the sensitivity to channel estimation errors. Besides, as we
have seen in Fig. 1.6, spread-spectrum achieves a higher aggregation of traffic
with lower interference fluctuations which is favorable for RA.](https://image.slidesharecdn.com/resolutionforrandomaccessschemes-220518201640-39d36c53/85/Resolution-for-Random-Access-Schemes-pdf-46-320.jpg)
![2.2. ASYNCHRONOUS RANDOM ACCESS 33
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2
0
0.2
0.4
0.6
0.8
1
Average MAC Channel Load − G [bits/symbol]
Throughput
−
T
[bits/symbol]
ACRDA Simulation
ACRDA Analytical
CRDSA Analytical
CRDSA Simulation
(a) Throughput
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2
10
−6
10
−5
10
−4
10
−3
10
−2
10
−1
10
0
Average MAC Channel Load − G [bits/symbol]
Packet
Loss
Ratio
[PLR]
CRDSA Simulated
ACRDA Simulated
ACRDA Analytical
CRDSA Analytical
(b) PLR
Figure 2.8: Simulation and analytical ACRDA and CRDSA performance for
Nrep = 2, Nslots=100 (simulations), QPSK modulation, 3GPP FEC ρ = 1/3,
packet block size 100 bits, Es/N0 = 10 dB in the presence of no packets power
imbalance and Poisson traffic, window size W = 3 virtual frames and a window
step ∆W = 0.15 [61].](https://image.slidesharecdn.com/resolutionforrandomaccessschemes-220518201640-39d36c53/85/Resolution-for-Random-Access-Schemes-pdf-47-320.jpg)
![34 CHAPTER 2. THE WORK
0 0.5 1 1.5 2 2.5
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
Average MAC Channel Load − G [bits/s/Hz]
Throughput
−
T
[bits/s/Hz] E−SSA σ=0 dB
E−SSA σ=3 dB
ACRDA σ=0 dB
ACRDA σ=3 dB
(a) Throughput
1 1.2 1.4 1.6 1.8 2
10
−6
10
−5
10
−4
10
−3
10
−2
10
−1
10
0
Average MAC Channel Load − G [bits/s/Hz]
Packet
Loss
Ratio
(PLR)
E−SSA σ=0 dB
E−SSA σ=3 dB
ACRDA σ=0 dB
ACRDA σ=3 dB
(b) PLR
Figure 2.9: Simulated E-SSA performance with BPSK modulation, 3GPP FEC
ρ=1/3, packet block size 100 bits, spreading factor 256 vs. ACRDA performance
for Nrep = 2, Nslots=100, QPSK modulation, 3GPP FEC ρ=1/3, packet block
size 100 bits, Es/N0 = 10 dB, window size W = 3 virtual frames and a window
step ∆W = 0.15, with and without power imbalance and Poisson traffic.](https://image.slidesharecdn.com/resolutionforrandomaccessschemes-220518201640-39d36c53/85/Resolution-for-Random-Access-Schemes-pdf-48-320.jpg)
![Chapter 3
Outcomes
The contributions from the author in this PhD thesis have derived in a number of
prototyping activities, commercial products and telecommunications standards.
They are briefly presented in this chapter.
3.1 Prototypes
A real-time prototype of a CRDSA modulator and demodulator has been de-
veloped by Space Engineering (Italy) in 2011 in the frame of an ESA contract
[63, 64]. Fig. 3.1 compares the measured results in the test bed against those
obtained by simulations. We can see that experimental results are very close to
the simulation findings assuming an ideal CRDSA demodulator.
A reference implementation of the E-SSA technique based on a software
defined radio (SDR) platform was developed by MBI (Italy) in 2012 under Eu-
telsat’s specifications [65, 66, 67]. The MBI prototype allowed to fully verify
the E-SSA performance in a real-time environment including preamble detec-
tion, channel estimation, interference cancellation and frequency errors. We can
see in Fig. 3.2 that the measured performance are very close to the simulated
ones. The PLR remains very low until almost the peak MAC throughput and,
as expected, power imbalance improves the E-SSA performance.
A second E-SSA prototype was developed in 2014 in the frame of an ESA
project, called ANTARES [68], aiming at the development of a new purpose-built
satellite communications system, including low-cost airborne terminals and a
new open satellite communication protocol for Air Traffic Management (ATM).
The new proposed communication protocol is based on E-SSA in the the re-
turn link (aircraft to hub) and is implemented in 200 kHz channels in L-band
(1-2 GHz) with a chip rate of 160 kchip/s and a spreading factor SF = 16 or
4 [69, 70]. The performance of the new proposed telecommunication protocol
have been verified means of an end-to-end Verification Test Bed (see Fig. 3.3).
Laboratory measurements with the verification testbed confirm that a through-
put above 1 bit/chip can be achieved with spreading factor SF = 16 (see Fig.
3.4) in the aeronautical channel with a target PLR = 10−3. Slightly lower spec-
tral efficiency, around 0.8 bits/s/Hz, is achieved with SF = 4 (see Fig. 3.5).
35](https://image.slidesharecdn.com/resolutionforrandomaccessschemes-220518201640-39d36c53/85/Resolution-for-Random-Access-Schemes-pdf-49-320.jpg)
![3.1. PROTOTYPES 37
0 0.5 1 1.5 2 2.5
10
-7
10
-6
10
-5
10
-4
10
-3
10
-2
10
-1
10
0
MAC load (bit/s/Hz)
Packet
Loss
Ratio
(PLR)
!=0dB, analitycal
!=2dB, analytical
!=3dB, analytical
!=0dB, simulation
!=2dB, simulation
!=3dB, simulation
!=0dB, laboratory
!=2dB, laboratory
!=3dB, laboratory
Figure 3.2: PLR vs. MAC loads and diffrent values of power imbalance standard
deviation (C/N=-16 dB) [65].](https://image.slidesharecdn.com/resolutionforrandomaccessschemes-220518201640-39d36c53/85/Resolution-for-Random-Access-Schemes-pdf-51-320.jpg)
![38 CHAPTER 3. OUTCOMES
It is also important to highlight the very good match between simulation and
measurement results.
An ACRDA prototype is currently under development by DLR (Germany)
under the frame of an ESA contract [71]. Measurement results will be available
in 2016.
In chapter 2 from [72], Riccardo de Gaundezi and Gennaro Gallinaro com-
pare the simulated CRDSA and E-SSA aggregated throughput versus the uncon-
strained and constrained in the used code rate and modulation. The simulated
CRDSA throughput with 3 replicas, a PLR=10−3, QPSK modulation and 3GPP
turbo code with 100 bits block size is compared to the unconstrained and QPSK
with FEC 100 bit block size constrained capacity bounds. In case of CRDSA
the packets have been using a uniform in dB random packet power distribution
with a maximum value of Eb/N0 of 6, 9, 12 and 20 dB respectively. It can be
observed that the simulated CRDSA throughput is quite close to the modula-
tion and FEC constrained capacity at medium values of Eb/N0. The distance is
increasing for large values of Eb/N0 possibly due to the fixed modulation format
(QPSK) and code rate (1/3) adopted for CRDSA. For E-SSA, with an average
SNR after de-spreading of 10 dB and using a uniform in dB random packet
power distribution with a range of 6 dB, an unconstrained capacity bound of
2.5 bits/s/Hz and a constrained capacity bound with BPSK, spreading factor
SF = 16 and FEC coding rate ρ = 1/3 of 2 bits/s/Hz are derived. These results
show that the simulated E-SSA aggregated throughput of 2 bit/s/Hz presented
in Sect. 2.2.1 match the modulation and FEC constrained capacity.
3.2 Standards and Systems
In the last years, a number of satellite communication standards have been
developed integrating the most recent RA techniques. In the broadband access
domain, the second generation of the DVB-RCS standard for interactive satellite
systems [73], has added CRDSA as an optional feature, which is particularly
advantageous for the SCADA and consumer profiles.
In the mobile context, the inclusion of the low-latency profile in the DVB-SH
standard for satellite services to handheld [74], enabled the possibility to add
a return link (earth station to hub), which is enhancing the DVB-SH standard
for the support of interactive services. The ETSI S-band Mobile Interactive
Multimedia (S-MIM) standard [75, 65] has adopted the E-SSA technique for the
asynchronous access.
The E-SSA RA technology is finding commercial application in the recently
deployed Eutelsat Broadcast Interactive System (EBIS) (see Fig. 3.6). The
core of the EBIS system is the Fixed Satellite Interactive Multimedia (F-SIM)
protocol [76]. The F-SIM protocol is implemented by the Smart LNB [77] on
the earth stations. The Smart LNB implements conventional DVB-S2 on the
forward link (hub to earth station) [78] and a proprietary evolution of the S-
MIM standard [75] in the return link (earth station to hub). The EBIS system
is already deployed in Ka-band on Eutelsat Ka-Sat satellite, as well as in Ku-
band on different Eutelsat satellites.](https://image.slidesharecdn.com/resolutionforrandomaccessschemes-220518201640-39d36c53/85/Resolution-for-Random-Access-Schemes-pdf-52-320.jpg)
![42 CHAPTER 3. OUTCOMES
Figure 3.6: Eutelsat Broadcast Interactive System [76].
3.3 Example of a M2M satellite network
In this section, a realistic example of an M2M satellite network is described with
the aim of describing the integration of the new proposed RA schemes with the
upper layer protocols and to perform a preliminary system dimensioning exercise
based on the traffic characteristics.
The selected scenario, is an aeronautical satellite network to track and mon-
itor airplanes flying over oceanic airspace (e.g. over the Atlantic). The system
uses the L-band (1-2 GHz) for radio links between the airplanes and the satel-
lite, with an uplink frequency of 1.55 GHz (aircraft to satellite) and a downlink
frequency of 1.65 GHz (satellite to aircraft). Frequency channels on uplink and
downlink have a bandwidth of 200 kHz. Aircrafts are equipped with a low gain
antenna and low power satellite communications terminal. A hub located on
ground communicates with the satellite using a frequency band allocated to
Fixed Satellite Services (e.g. in C, Ku or Ka-band). The satellite relays the sig-
nals received from the hub to the airplanes (forward link) as well as the signals
received from the airplanes to the hub (return link). The hub is connected via a
ground network to the to the Air Traffic Control (ATC) centre. A bidirectional
communication link is established between the ATC centre and the airplanes.
The network architecture is depicted in Fig. 3.7.
Airplanes transmit position reports periodically, and also triggered by an
event on-board or upon request from the ATC centre. On average, a message
is generated every 40 seconds per aircraft with an average length of 250 bytes,
resulting in an average bit rate per aircraft of 50 bits/s. Headers from network](https://image.slidesharecdn.com/resolutionforrandomaccessschemes-220518201640-39d36c53/85/Resolution-for-Random-Access-Schemes-pdf-56-320.jpg)
![3.3. EXAMPLE OF A M2M SATELLITE NETWORK 43
Land
Ocean
Satellite HUB
Air Traffic Control
L-band
C-, Ku- or
Ka-band
Figure 3.7: Aeronautical Satellite Network for Air Traffic Management.
protocols are compressed down to 4-7 bytes, e.g. using ROHC [79, 80], which
allows an efficient use of bandwidth on low- and medium-speed links. At the
link layer, the Generic Stream Encapsulation (GSE) protocol defined in [81] and
adapted to the aeronautical environment in [69]. This encapsulation protocol
is characterized by its flexibility and low overhead in the range of 3-4%. If we
account for the compressed network headers and encapsulation overhead, the
resulting average information bit rate per aircraft at physical layer is 54 bits/s.
We have seen in Sect. 3.1 that the ANTARES project proposes an air in-
terface that can achieve spectral efficiencies in the return link above 1 bit/chip
for a target PLR = 10−3 by using the E-SSA technique. This means that in
a 160 kchip/s channel (200 kHz bandwidth) we can track and monitor ≈ 3000
airplanes flying simultaneously over the oceanic airspace. Today, there are less
of 3000 aircrafts flying over the Atlantic ocean over a whole day. At a bit rate
of 3.3kbit/s per aircraft, the average transmission delay for each message will be
650ms, and the message will be received by the hub in less than 1s (if we account
for the propagation delay). In total, the end-to-end message transmission delay
from aircraft to Air Traffic Control centre will be < 1.5s, if we account for hub
message decoding time and ground network delay from hub to satellite control
centre.
These are remarkable performance results with a highly aggregated through-
put, low PLR and low end-to-end delay, if we consider that a very large network
population is sharing a communication channel with a very large propagation
delay.](https://image.slidesharecdn.com/resolutionforrandomaccessschemes-220518201640-39d36c53/85/Resolution-for-Random-Access-Schemes-pdf-57-320.jpg)
![195
account the fact that thanks to powerful physical layer coding, packets may be decoded
with non zero probability also in the presence of collisions. The model also considers
the effect of slot time diversity in case more than one packet replica is transmitted
in the same frame, as it is the case for DSA. Two types of IC strategies have been
modeled: successive interference cancellation over one slot position (CRDSA) or one
window position (E-SSA and ACRDA), and iterative interference cancellation across the
slots due to successful reception of packet replicas in other slots (CRDSA and ACRDA).
The new model have been used to perform sensitivity analyses against a number of
design parameters for the new RA schemes presented in this thesis, and to compare the
performance improvement against conventional RA schemes. The good match between
analytical and simulation results confirm the validity of the models, and make them
extensible for future use in new or evolved RA schemes.
The new RA schemes have triggered the interest from industry who have developed
real time laboratory prototypes [63, 67, 71] and validated the simulation results. In ad-
dition, CRDSA and E-SSA have been incorporated into two satellite telecommunication
standards confirming the interest from industry [73, 75]. Finally, we conclude this thesis
by mentioning that the E-SSA RA technology is finding commercial application in the
recently deployed Eutelsat Broadcast Interactive System [76].](https://image.slidesharecdn.com/resolutionforrandomaccessschemes-220518201640-39d36c53/85/Resolution-for-Random-Access-Schemes-pdf-85-320.jpg)
![Chapter 12
Future work
The development of the new RA techniques presented in this thesis has opened up a
new field of research. In this chapter some of the key open areas of research related to
high performance satellite RA are listed.
The choice of the physical layer Forward Error Correcting (FEC) scheme for CRDSA
is an important design driver. In CRDSA we are not operating in AWGN-like condi-
tions, but in the presence of heavy co-channel interference the FEC collision resolution
capability is important. Under these conditions of heavy multiple access interference, it
is important to choose a FEC code that is able to recover a few packets and will then
trigger the iterative IC process. Therefore, lower coding rates, although apparently re-
ducing the individual packet bit rate, may be important to enhance the aggregated RA
scheme performance. The final choice of coding rate is a tradeoff between increased
overhead and RA performance improvement and has been preliminary studied in [60].
Regarding the FEC block size, it is preferable to have a small block size as this provides
a smaller PLR in the lower signal-to-noise ratio (SNR) range [60]. Again, the reasoning
is that successful detection of few packets under heavy co-channel interference conditions
will be able to trigger the iterative IC process and eventually recover all packets from the
RA frame. As a general rule, low coding rates (≤ 1/2) and small packet sizes (≤ 1000
bits) are recommended for CRDSA and in general, all RA schemes implementing IC.
Another interesting subject of research is related to reduction of the transmitter
power consumption and cost. The use of constant envelope modulation like Continuous
Phase Modulation (CPM) can bring several advantages, as it allows to replace the
transmitter frequency up-conversion unit with a cheaper nonlinear frequency multiplier.
Furthermore the use of CPM modulation allows to operate the transmitter amplifier in
highly nonlinear mode with consequent reduction of the DC power requirements. Past
work has shown the possibility to combine CPM with DS-SS [149]. Although CPM
has been used in commercial S-ALOHA and SSA systems such as Sat3Play [150] and
ArcLight [151], its use for iterative SIC based RA is still under investigation.
While for SSA and E-SSA the use of low-order modulation such as BPSK with low
coding rate is close to optimum, recent investigations [152] indicate that for non-SS
RA there may be some throughput advantage using modulation orders above QPSK
when the SNR is above a certain value. Using QPSK in E-SSA does not bring any
throughput advantage, but just an increase on the sensitivity to channel estimation phase
errors. The use of a pilot in quadrature allows to have a continuous phase estimation
which provides robustness in case of channels with phase noise and/or fading (mobile
environment). However, concerning CRDSA, the use of QPSK instead of BPSK is
related to the achievable spectral efficiency considering the TDMA type of access scheme.
The idea is to use an energy efficient signaling scheme like QPSK (or a higher order
197](https://image.slidesharecdn.com/resolutionforrandomaccessschemes-220518201640-39d36c53/85/Resolution-for-Random-Access-Schemes-pdf-87-320.jpg)
![198 CHAPTER 12. FUTURE WORK
modulation) with low rate FEC to minimize the power of the transmitter while getting
very good aggregate RA throughput.
When the RA network allows to exploit the incoming packet power imbalance, this
power distribution can be optimized to enhance the system throughput. While for E-
SSA RA some semi-analytical optimization has been derived in [153], a similar effort
for non spread spectrum RA is still undergoing. Initial findings indicate that also for
CRDSA and ACRDA a uniform in dB incoming packet distribution can improve the
throughput performance.
In systems with a high SNR, but reduced packet power dynamic range, MMSE
preprocessing may be implemented to enhance the SNIR of the received signal before
the signal demodulation. The MMSE detector can exploit the knowledge of the signal
structure to compute the matrix filter. Preliminary analyses and simulation results on
a newly proposed ME-SSA scheme can be found in [154].
A major drawback of TDMA is its high peak transmit power requirement resulting
from the fact that each terminal is transmitting on a small fraction (1/Nslots) of the
frame duration. Since the PLR of a system is directly related to the ratio between
the bit energy and the noise floor level (i.e. the Eb/N0), TDMA needs a (peak) power
transmission Nslots times higher than FDMA or CDMA. In thin route satellite commu-
nications networks Multi-Frequency TDMA (MF-TDMA) was introduced to mitigate
this problem. The multi-frequency concept can be easily applied to CRDSA or ACRDA
leading to MF-CRDSA [58]. The peak power gain reduction as for MF-CRDSA will be
proportional to the ratio of frequency to time slots of the system. Besides, using the
frequency domain will reduce the RA frame duration thus reducing the end-to-end delay
in CRDSA. The multi-frequency concept can be also applied in ACRDA to reduce the
terminal peak power requirements.
The exploitation in ACRDA of techniques such as soft combining of packet replicas
proposed in [140] or the combining of the best replica packets chunks to generate a better
packet of interest as suggested in [143] or the Coded Slotted ALOHA soft combining
scheme [136] represent potential area of future investigation.
Similarly one can consider the techniques proposed in [139], [140] to pseudo-randomize
the slot positions as a function of the transmitted data to remove the replica signaling
overhead. This is particularly interesting in case small packets have to be transmitted.
The issue of advanced Spread Spectrum RA demodulation in case of multi-beam
satellite networks has been recently tackled in [155]. It has been shown that for a
mobile multi-beam system exploiting the S-MIM standard further RA capacity gain
can be achieved using full frequency reuse among the beams and joint beam processing
at the gateway. In particular, two different multi-beam joint processing techniques
have been investigated by simulation: basic (decoded packets are send to upper layers
belonging to other beams) and enhanced cooperation (the decoded data is sent to the
co-frequency demodulators on top the basic cooperation). The latter approach, although
more complex to implement, brings extra gain on top of the basic cooperation. More
work in this area is required to assess the practical feasibility and gains achievable
exploiting joint beam processing.
In this thesis we have shown that RA alone can achieve throughput performance
truly comparable with DAMA schemes with advantages in terms of transmission latency
and signaling reduction. Furthermore, the presented advanced RA techniques are easily
scalable to very large networks. However, there may be applications where the type of
traffic can be both bursty or bulky. In this case the combination of high performance RA
schemes with DAMA can represent the best solution to cope with this very time variant
traffic conditions. More work is required to investigate the best way to adaptively switch
from the new efficient and reliable RA to DAMA type of access schemes [144, 156].
Finally, the techniques presented in this thesis can also be adopted in terrestrial](https://image.slidesharecdn.com/resolutionforrandomaccessschemes-220518201640-39d36c53/85/Resolution-for-Random-Access-Schemes-pdf-88-320.jpg)
![199
wireless networks to improve their performance in terms of capacity and energy con-
sumption. Examples systems for application are Radio Frequency Identification (RFID)
systems [139, 140] and M2M networks [157, 158].](https://image.slidesharecdn.com/resolutionforrandomaccessschemes-220518201640-39d36c53/85/Resolution-for-Random-Access-Schemes-pdf-89-320.jpg)
![Bibliography
[1] M. Williamson, Spacecraft Technology: The Early Years, The Institution of Engi-
neering and Technology, 2006.
[2] K.M.S. Murthy, J. Alan, J. Barry, B.G. Evans, N. Miller, R. Mullinax, P. Noble, B.
O’Neal, J.J. Sanchez, N. Seshagiri, D. Shanley, J. Stratigos, J.W. Warner, VSAT
user network examples, IEEE Communications Magazine, Vol. 27, Issue 5, pp. 50-
57, May 1989.
[3] B. Arbesser-Rastburg, R. Bellini, F. Coromina, R. De Gaudenzi, O. del Rio Her-
rero, M. Hollreiser, R. Rinaldo, P. Rinous, A. Roederer, R&D directions for next
generation broadband multimedia systems: an ESA perspective, 20th AIAA Interna-
tional Communications Satellite Systems Conference (ICSSC), Montreal, Canada,
May 13-15, 2002.
[4] Satellite Industry Association State of Satellite Industry Report, September 2014.
[5] Euroconsult, High throughput satellites: on course for new horizons, 2014 Edition.
[6] Northern Sky Research, HTS Revenues to Overtake FSS in Battleground Applica-
tions, http://www.nsr.com/news-resources/nsr-in-the-press/nsr-press-releases/hts-
revenues-to-overtake-fss-in-battleground-applications/.
[7] Via Satellite, The Continuing Momentum of M2M/IoT via Satellite,
http://www.satellitetoday.com/publications/2015/01/20/the-continuing-
momentum-of-m2miot-via-satellite/.
[8] G. Maral, M. Bousquet, Satellite Communicaitons Systems, Fifth Edition, John
Wiley & Sons Ltd, 2009.
[9] ETSI EN 300 421 Ver. 1.1.2, Digital Video Broadcasting (DVB); Framing structure,
channel coding and modulation for 11/12 GHz satellite services, August 1997.
[10] ETSI EN 302 307-1 Ver. 1.4.1, Digital Video Broadcasting (DVB); Second gener-
ation framing structure, channel coding and modulation systems for Broadcasting,
Interactive Services, News Gathering and other broadband satellite applications;
Part 1: DVB-S2, November 2014.
[11] G. Maral, VSAT Networks, Second Edition, John Wiley & Sons Ltd, 2003.
[12] A.H. Rana, J.S. McCoskey, W.A. Check, VSAT technology, trends, and applica-
tions, Proceedings of the IEEE, Vol. 78, Issue 7, pp. 1087 - 1095, July 1990.
[13] D. Chakraborty, VSAT Communications Networks - An overview, IEEE Commu-
nications Magazine, Vol. 26, No. 5, pp. 10-24, May 1988.
[14] F.A. Tobagi, Multiaccess Protocols in Packet Communication Systems, IEEE Trans-
actions on Communications, Vol. COM-28, No. 4, pp. 468-488, April 1980.
201](https://image.slidesharecdn.com/resolutionforrandomaccessschemes-220518201640-39d36c53/85/Resolution-for-Random-Access-Schemes-pdf-91-320.jpg)
![202 BIBLIOGRAPHY
[15] C. Retnadhas, Satellite Multiple Access Protocols, IEEE Communications Maga-
zine, Vol. 18, No. 5, pp. 16-20, September 1980.
[16] C.J. Wolejsza, D. Taylor, M. Grossman, W.P. Osborne, Multiple Access Protocols
for Data Communications via VSAT Networks, IEEE Communications Magazine,
Vol. 25, No. 7, July 1987.
[17] D. Raychaudhuri, K. Joseph, Channel Access Protocols for Ku-band VSAT Net-
works: a comparative evaluation, IEEE Communications Magazine, Vol. 26, No. 5,
pp. 34-44, May 1988.
[18] H. Peyravi, Medium Access Control Protocol Performance in Satellite Communica-
tions, IEEE Communications Magazine, Vol. 37, No. 3, pp. 62-71, March 1999.
[19] P.M. Feldman, An Overview and Comparison of Demand Assignment Multiple Ac-
cess (DAMA) Concepts for Satellite Communications Networks, RAND corpora-
tion, 1996.
[20] J.G. Puente, A PCM-FDMA demand assignment satellite multiple access experi-
ment, AIAA 2nd Communications Satellite Systems Conference, paper No. 68-451,
8-10 April 1968, San Francisco (USA).
[21] G. Frenkel, The grade of service in multiple-access satellite communications systems
with demand assignments, IEEE Transactions on Communications, Vol. 22, No. 10,
pp. 1681-1685, October 1974.
[22] W.E. Leland, On the self-similar nature of Ethernet traffic (Extended version),
IEEE/ACM Transactions on Networking, Vol. 2, No. 1, February 1994.
[23] V. Paxson, S. Floyd, Wide Area Traffic: the failure of Poisson modeling,
IEEE/ACM Transactions on Networking, Vol. 3, No. 3, June 1995.
[24] I. Jacobs, L.N. Lee, A. Viterbi, R. Binder, R. Bressler, N.T. Hsu, R. Weissler,
CPODA - A demand Assignment protocol for Satnet, in Proceeding of the fifth
symposium on data communications (SIGCOMM’77), pp. 2.5-2.9, 1977.
[25] T. Le-Ngoc, S.V. Krishnamurthy, Performance of combined free/demand assign-
ment multiple-access schemes in satellite communications, Wiley International jour-
nal of satellite communications, Vol. 14, No. 1, pp. 11-21, January 1996.
[26] T. Le-Ngoc and I.M. Jahangir, Performance analysis of CFDAMA-PB protocol for
packet satellite communications, IEEE Transactions on Communications, Vol. 46,
No. 9, pp. 1206-1214, September 1998.
[27] O. del Rı́o Herrero, G. Foti and G. Gallinaro, Spread-spectrum techniques for the
provision of packet access on the reverse link of next-generation broadband multi-
media satellite systems, IEEE Journal on Selected Areas in Communications, Vol.
22, Issue 3, pp. 574 - 583, April 2004
[28] Z. Jiang, Y. Li, V.C.M. Leung, A Predictive Demand Assignment Multiple Access
Protocol for Broadband Satellite Networks Supporting Internet Applications, IEEE
International Conference on Communications (ICC), pp. 2973 - 2977, Vol.5, 28
April - 2 May 2002, New York (USA)
[29] M. Hasegawa, G. Wu, M. Mizuno, Applications of nonlinear prediction methods to
the Internet traffic, Proc. International Symposium on Circuits and Systems, vol.
3, pp. 169-172, May 2001, Sydney (Australia).](https://image.slidesharecdn.com/resolutionforrandomaccessschemes-220518201640-39d36c53/85/Resolution-for-Random-Access-Schemes-pdf-92-320.jpg)
![BIBLIOGRAPHY 203
[30] H.W. Lee and J. W.Mark, Combined random/reservation access for packet switched
transmission over a satellite with on-board processing: Part I Global beam satellite,
IEEE Transactions on Communications, Vol. COM-31, No. 10, pp. 1161-1171, Oc-
tober 1983.
[31] D. M. Chitre, VSAT Networks: Architectures, Protocols, and Management, IEEE
Communications Magazine, July 1988, Vol. 26, No. 7, pp. 28-38, July 1988.
[32] D. Bertsekas and R. Gallager, Data Networks, Second Edition, Prentice Hall, 1992.
[33] S. Keshav, An Engineering Approach to Computer Networking, Addison-Wesley
Professional, 1997.
[34] L. Kleinrock and F.A. Tobagi, Packet switching in radio channels: part I - Car-
rier sense multiple access modes and their throughput-delay characteristics, IEEE
Transactions on Communications, Vol. COM-23, No. 12, pp. 1400-1416, December
1975.
[35] R. M. Metcalfe and D.R. Boggs, ETHERNET: Distributed packet switching for
local computer networks, Communications of the ACM magazine, Vol. 19, No. 7,
pp. 395-404, July 1976.
[36] F.A. Tobagi and L. Kleinrock, Packet Switching in Radio Channels: PartII - The
Hidden Terminal Problem in Carrier Sense Multiple-Access and the Busy-Tone
Solution, IEEE Transactions on Communications, Vol. COM-23, No. 12, December
1975.
[37] IEEE Std 802.3-2005, IEEE Standard for Information technology; Telecommunica-
tions and information exchange between systems local and metropolitan area net-
works; Specific requirements Part 3: Carrier sense multiple access with collision
detection (CSMA/CD) access method and physical layer specifications, December
2005.
[38] IEEE Std 802.11-2012, IEEE Standard for Information technology; Telecommu-
nications and information exchange between systems local and metropolitan area
networks; Specific requirements Part 11: Wireless LAN Medium Access Control
(MAC) and Physical Layer (PHY) Specifications, March 2012.
[39] P. Karn, MACA - A New Channel Access Method for Packet Radio, Proceedings
of the ARRL/CRRL Amateur Radio 9th Computer Networking Conference, pp.
134-140, 22 September 1990, London and Ontario (Canada).
[40] V. Bharghavan, A. Demers, S. Shenker, L. Zhang, MACAW: A Media Access Pro-
tocol for Wireless LAN’s, ACM SIGCOMM, 1994.
[41] C. Fullmer, J.J. Garcia-Luna-Aceves, Floor Acquisition Multiple Access (FAMA)
for packet radio networks, Computer Communication Review, Vol. 25, No. 4, ACM
SIGCOMM ’95, Cambridge, MA, USA, 28 Aug.-1 Sept. 1995.
[42] 3GPP TS 25.214 v3.12.0, Physical Layer Procedures (FDD); Release 1999, March
2003.
[43] N. Abramson, The ALOHA system - Another alternative for computer communi-
cations, in 1970 Fall Joint Comput. Conf. AFIPS Conf. Proc., vol. 37, pp. 281-285,
AFIPS Press, Montvale (NJ), 1970.
[44] N. Abramson, The Throughput of Packet Broadcasting Channels, IEEE Transac-
tions on Communications, Vol. COM-25, No. 1, pp. 117-128, January 1977.](https://image.slidesharecdn.com/resolutionforrandomaccessschemes-220518201640-39d36c53/85/Resolution-for-Random-Access-Schemes-pdf-93-320.jpg)
![204 BIBLIOGRAPHY
[45] L. G. Roberts, ALOHA Packet Systems with and Without Slots and Capture,
Arpanet Satellite System Note 8 (NIC Document #11290), 26 Jun. 1972.
[46] G.L. Choudhury and S.S. Rappaport, Diversity ALOHA - A Random Access
Scheme for Satellite Communications, IEEE Transactions on Communications, Vol.
COM-31, pp. 450-457, March 1983.
[47] O. del Rı́o Herrero and R. De Gaudenzi, Generalized Analytical Framework for the
Performance Assessment of Slotted Random Access Protocols, IEEE Transactions
on Wireless Communications, Vol. 13, No. 2, pp. 809-821, February 2014.
[48] N. Abramson, Multiple Access in Wireless Digital Networks, IEEE Proceedings,
Vol. 82, No. 9, pp. 1360-1370, September 1994.
[49] C. Pateros, Novel direct sequence spread spectrum multiple access technique, in
MILCOM 2000, 21st Century Military Communications Conference Proceedings,
Vol. 2, pp. 564-568, Los Angeles (CA) Oct. 22-25, 2000.
[50] O. del Rı́o Herrero and R. De Gaudenzi, High Efficiency Satellite Multiple Ac-
cess Scheme for Machine-to-Machine Communications, IEEE Transactions on
Aerospace and Electronic Systems, Vol. 48, No. 4, pp. 2961-2989, October 2012.
[51] ETSI EN 301 790 v1.5.1, Digital Video Broadcasting (DVB); Interaction channel
for satellite distribution systems, May 2009.
[52] Telecommunication Industry Association TIA-1008-A, IP Over Satellite, May. 2006.
[53] S. Verdu, Multiuser detection, Cambridge Univ. Press, 1998.
[54] C. C. Lee, Random signal levels for channel access in packet broadcast networks,
IEEE J. Select. Areas Commun., Vol. 5, No. 6, pp. 1026-1034, July 1987.
[55] I. Cidon, J. Kodesh and M. Sidi, em Erasure, capture, and random power level
selection in multiple access systems, IEEE Transactions on Communications, Vol.
36, No. 3, pp. 263-271, March 1988.
[56] R. O. LaMaire, A. Krishna and M. Zorzi, On the randomization of transmitter power
levels to increase throughput in multiple access radio systems, Journal Wireless
Networks, Vol. 4, No. 3, pp. 263 - 277, March 1998, Kluwer Academic Publishers
Hingham, MA (USA).
[57] E. Casini, R. De Gaudenzi, O. del Rı́o Herrero, Contention Resolution Diversity
Slotted Aloha (CRDSA): an Enhanced Random Access Scheme for Satellite Access
Packet Networks, IEEE Transactions on Wireless Communications, Vol. 6, No. 4,
pp. 1408-1419, Apr. 2007.
[58] E. Casini, O. del Rı́o Herrero, R. De Gaudenzi, D.M.E. Delaruelle, J.P. Choffray,
Transmission de donnees par paquets a travers un canal de transmission partage
par plusierus utilisateurs, European Patent EP1686746 (B1), 30 January 2006.
[59] O. del Rı́o Herrero, R. De Gaudenzi, Methods, apparatuses and system for asyn-
chronous spread-spectrum communications, U.S. Patent 7,990,874 B2, August 2,
2011.
[60] O. del Rı́o Herrero, R. De Gaudenzi, J. Pijoan Vidal, Design Guidelines for Ad-
vanced Random Access Protocols, 30st AIAA International Communications Satel-
lite Systems Conference (ICSSC), 24-27 Sep. 2012, Ottawa (Canada).
[61] R. De Gaudenzi, O. del Rı́o Herrero, G. Acar, E. Garrido Barrabés, Asynchronous
Contention Resolution Diversity Aloha: Making CRDSA Truly Asynchronous,
IEEE Transactions on Wireless Communications, Vol. 13, No. 11, pp. 6193-6206,
November 2014.](https://image.slidesharecdn.com/resolutionforrandomaccessschemes-220518201640-39d36c53/85/Resolution-for-Random-Access-Schemes-pdf-94-320.jpg)
![BIBLIOGRAPHY 205
[62] R. De Gaudenzi, O. del Rı́o Herrero, Method and apparatus for transmitting data
packets over a transmission channel shared by a plurality of users, World Intellectual
Property Organization Patent WO2015000518 (A1), 3 July 2013.
[63] Space Engineering, Advanced Modem Prototype for Interactive Satellite Terminals
(AMPIST), ESA Contract No. 20668/07/NL/JD, https://artes.esa.int/advanced-
modem-prototype-interactive-satellite-terminals-ampist
[64] R.Romanato, E. Rossini, G. Chiassarini, A. Masci, D. Silvi, P. Burzigotti, D. Gi-
ancristofaro, AMPIST: An Advanced Modem Prototype For Interactive Satellite
Terminals with Contention Resolution Diversity Slotted ALOHA (CRDSA) Capa-
bilities, in 17th Ka and Broadband Communications, Navigation and Earth Obser-
vation Conference, 3-5 October 2011, Palermo (Italy)
[65] S. Scalise, C.P. Niebla, R. De Gaudenzi, O. del Rio Herrero, D. Finocchiaro, A.
Arcidiacono, S-MIM: a novel radio interface for efficient messaging services over
satellite, IEEE Communications Magazine, Vol. 51 , No. 3, pp. 119-125, March
2013.
[66] F. Basile, G. Mendola, Satellite Hub Communication System GPU Based, GPU
Technology Conference, 14-17 May 2012, San Jose (CA).
[67] M. Andreacci and G. Mendola, F. Collard, D. Finocchiaro, A. Recchia, Enhanced
spread spectrum ALOHA demodulator implementation, laboratory tests and satellite
validation, Wiley International Journal on Satellite Communications and Network-
ing, Special Issue on S-band Mobile Interactive Multimedia, Vol. 32, No. 6, Page
521-533, November/December 2014.
[68] AeroNauTicAl REsources Satellite based (ANTARES), ESA Contract
22851/09/NL/AD, https://artes.esa.int/projects/antares
[69] ANTARES Communication Standard Technical Specifications,
https://artes.esa.int/news/iris-antares-study-updates-draft-satellite-
communications-standard
[70] J. Batlle, I. Vega, J. Vila, D. Fernandez, L. Albiol, J.M. Cebrian, C. Ciancarelli,
P. Burzigotti, New satellite-based communication standard for Air Traffic Manage-
ment, in 19th Ka and Broadband Communications, Navigation and Earth Obser-
vation Conference and 31st AIAA International Communications Satellite Systems
Conference (ICSSC), 14-17 October 2013, Florence (Italy).
[71] Prototypes for Advanced Access Schemes, ESA Contract 4000110938/14/NL/CLP,
2014.
[72] O. del Rı́o Herrero, R. De Gaudenzi, G. Gallinaro, High-performance random ac-
cess schemes, in Co-operative and cognitive satellite systems, S. Chatzinotas, B.
Ottersten, R. De Gaudenzi, Ed. Academic Press, 2015, pp. 35-82.
[73] ETSI EN 301 542-2 v1.1.1, Digital Video Broadcasting (DVB); Second Generation
DVB Interactive Satellite System (DVB-RCS2); Part 2: Lower Layers for Satellite
Standard, January 2012.
[74] ETSI EN 302 583 v1.1.1, Digital Video Broadcasting (DVB); Framing Structure,
channel coding and modulation for Satellite Services to Handheld devices (SH) below
3 GHz, March 2008.
[75] ETSI TS 102 721-3 v1.1.1, Satellite Earth Stations and Systems; Air Interface for S-
band Mobile Interactive Multimedia (S-MIM); Part 3: Physical Layer Specification,
Return Link Asynchronous Access, December 2011.](https://image.slidesharecdn.com/resolutionforrandomaccessschemes-220518201640-39d36c53/85/Resolution-for-Random-Access-Schemes-pdf-95-320.jpg)
![206 BIBLIOGRAPHY
[76] A. Arcidiacono, D. Finocchiaro, F. Collard, S. Scalise, F. Lazaro Blasco, R. De
Gaudenzi, S. Cioni, N. Alagha, M. Andrenacci, From S-MIM to F-SIM: Making
Satellite Interactivity Affordable at Ku and Ka-band, Accepted for Publication in
Wiley International Journal on Satellite Communications and Networking, special
issue on Ka-band satellite communications, 2015.
[77] Eutelsat, Eutelsats Smart LNB for Direct-to-Home antenna,
http://www.eutelsat.com/en/services/broadcast/direct-to-home/smart-LNB.html
[78] ETSI EN 302 207 v1.3.1, Digital Video Broadcasting (DVB); Second generation
framing structure, channel coding and modulation systems for Broadcasting, Inter-
active Services, News Gathering and other broadband satellite applications (DVB-
S2), March 2013.
[79] IETF RFC 5795, The RObust Header Compression (ROHC) Framework, March
2010.
[80] IETF RFC 3095, RObust Header Compression (ROHC): Framework and four pro-
files: RTP, UDP, ESP, and uncompressed, July 2001.
[81] J. Cantillo, B. Collini-Nocker, U. De Bie, O. del Rio, G. Fairhurst, A. Jahn R.
Rinaldo,GSE: A Flexible, yet Efficient Encapsulation for IP over DVB-S2 Con-
tinuous Generic Streams, International Journal of Satellite Communications and
Networking (IJSCN), vol. 26, Issue 3, pp. 231-250, May/Jun. 2008.
[82] T. Le-Ngoc, and J.I. Mohammed, Combined free/demand assignment multiple ac-
cess (CFDAMA) protocols for packet satellite communications; Universal Personal
Communications, in Conf. Rec. 2nd International Conference on Personal Com-
munications, 12-15 Oct. 1993, vol. 2, pp. 824 - 828.
[83] S. S. Lam, A carrier sense multiple access protocol for local networks, Computer
Networks, vol. 4, pp. 21-32, 1980.
[84] D. Raychaudhuri, Aloha with Multipacket Messages and ARQ-type Retransmission
Protocols-Throughput Analysis, IEEE Trans. on Comm., vol. COM-32, no. 2, pp.
148-154, Feb. 1984.
[85] D. Raychaudhuri, Stability, Throughput, and Delay of Asynchronous Selective Re-
ject Aloha, IEEE Trans. on Comm., Vol. 35, No. 7, pp. 767-772, July 1987.
[86] European Telecommunication Standardisation Institute (ETSI) TR 101 790 V1.2.1,
Digital Video Broadcasting (DVB); Interaction channel for Satellite Distribution
Systems; Guidelines for the use of EN 301 790, January 2003.
[87] P. Patel, and J. Holtzman, Analysis of a simple successive interference cancellation
scheme in a DS/CDMA system, IEEE Journal on Selected Areas in Comm., Vol.
12 , No. 5, pp. 796 - 807, June 1994.
[88] R. De Gaudenzi, F. Giannetti, and M. Luise, Signal Recognition and Signature Code
Acquisition in CDMA Receivers for Mobile Communications, IEEE Transactions on
Vehicular Tech., Vol. 47, No. 1, pp. 196 - 208, Feb. 1998.
[89] European Telecommunication Standardisation Institute (ETSI) TR 101 112 v3.2.0,
Universal Mobile Telecommunications System (UMTS); Selection procedures for the
choice of radio transmission technologies of the UMTS, April 1998.
[90] R. De Gaudenzi, O. del Rı́o Herrero, Advances in Random Access protocols for
satellite networks, International Workshop on Satellite and Space Communications
(IWSSC), pp. 331-336, 9-11 September 2009, Siena (Italy).](https://image.slidesharecdn.com/resolutionforrandomaccessschemes-220518201640-39d36c53/85/Resolution-for-Random-Access-Schemes-pdf-96-320.jpg)
![BIBLIOGRAPHY 207
[91] A.H. Abdelmonem, T.N. Saadawi, Performance analysis of spread spectrum packet
radio network with channel load sensing, IEEE Journal on Selected Areas in Com-
munications, Vol. 7, No. 1, January 1989, pp. 161-166.
[92] O. del Rı́o Herrero, R. De Gaudenzi, A High-Performance MAC Protocol for
Consumer Broadband Satellite Systems, Proceedings of 27th AIAA International
Communications Satellite Systems Conference )ICSSC), 1-4 June 2009, Edinburgh
(United Kingdom).
[93] K. S. Ghilousen, I. M. Jacobs, R. Padovani, A. J. Vitebi, L. A. Weaver, C. W.
Wheatley III, On the Capacity of a Cellular CDMA System, IEEE Transactions on
Vehicular Technology, Vol. 40. No. 2, pp. 303-312, May 1991.
[94] R. De Gaudenzi, F. Giannetti, Analysis of an Advanced Satellite Digital Audio
Broadcasting System and Complementary Terrestrial Gap-Filler Single Frequency
Network, IEEE Transactions on Vehicular Technology Vol. 42, No. 2, pp. 194-210,
May 1994.
[95] P. Cardieri, T. S. Rappaport, Statistics of the Sum of Lognormal Variables in Wire-
less Communications, Vehicular Technology Conference 2000, Proceedings of the
51st Vehicular Technology Conference Spring, Volume: 3, pp. 1823-1827, May 2000,
Tokyo (Japan).
[96] 3GPP TS 25.212 v3.11.0, Multiplexing and channel coding (FDD); Release 1999,
September 2002.
[97] A. Giorgetti, M. Chiani, Influence of Fading on the Gaussian Approximation for
BPSK and QPSK With Asynchronous Cochannel Interference, IEEE Transactions
on Wireless Comm., Vol. 4, No. 2, pp. 384-388, March 2005.
[98] J.G. Proakis, Digital Communications, Fourth Ed. New York, McGraw-Hill, 2001.
[99] D.H. Layer, Digital radio takes to the road, IEEE Spectrum, Vol. 38, No. 7, pp.
40-46, July 2001.
[100] C.Y. Yeun, Mobile TV technologies, in the Proc. of the 5th International Con-
ference on Information and Communications Technology (ICICT), pp. 2-9, 16-18
December 2007, Cairo (Egypt).
[101] 3GPP TR 22.868 v8.0.0, Study on Facilitating Machine to Machine Communica-
tion in 3GPP Systems; Release 8, March 2007.
[102] European Telecommunication Standardisation Institute (ETSI), Document Draft
ETSI TS 102 689 v0.3.1, Machine-to-Machine communications (M2M); M2M ser-
vice requirements, 2009.
[103] C. Pateros, Novel direct sequence spread spectrum multiple access technique, MIL-
COM 2000, 21st Century Military Communications Conference Proceedings, 22-25
October 2000, Los Angeles (USA)
[104] H. Holma, A. Toskala, WCDMA for UMTS, John Wiley & Sons, Ltd, 2005.
[105] T. Le-Ngoc, V. Leung, P. Takats and P. Garland, Interactive Multimedia Satellite
Access Communications, IEEE Communications Magazine, Vol. 41, No. 7, pp. 78-
85, July 2003.
[106] C.E. Wheatley III, R. Padovani, E. Zehavi, Fast forward link power control in a
code division multiple access system, United States Patent No. 5461639, 24 October
1995.](https://image.slidesharecdn.com/resolutionforrandomaccessschemes-220518201640-39d36c53/85/Resolution-for-Random-Access-Schemes-pdf-97-320.jpg)
![208 BIBLIOGRAPHY
[107] K.S. Gilhousen, R. Padovani, C.E. Wheatley III, Spread spectrum transmitter
power control method and system, United States Patent No. 5,257,283, 26 October
1993.
[108] A.M. Monk, L.B. Milstein, Open-loop power control error in a land mobile satellite
system, IEEE Journal on Selected Areas in Communications, Vol. 13, No. 2, pp.
205-212, February 1995.
[109] D. Boudreau, G. Caire, G. E. Corazza, R. De Gaudenzi, G. Gallinaro, M. Luglio,
R. Lyons, J. Romero-Garcia, A. Vernucci, H. Widmer, Wideband CDMA for the
Satellite Component ofUMTS/IMT-2000, IEEE Transactions on Vehicular Tech-
nology, Vol. 51, No. 2, pp 306-330, Mar. 2002.
[110] N. Abramson, Spread aloha CDMA data communications, United States Patent
No. 5537397, 16 July 1996.
[111] Xiang Feng, Yan Li, Guangguo Bi, A CDMA-slotted ALOHA broadband system
for multiservices, IEEE 1998 International Conference on Universal Personal Com-
munications (ICUPC’98), Vol. 2, pp. 1131-1135, 5-9 October 1998, Florence (Italy).
[112] Y. Tadokoro, H. Okada, T. Yamazato, A. Katayama, A. Ogawa, A new packet
detection scheme in CDMA unslotted ALOHA system with successive interference
cancellation, IEEE Global Telecommunications Conference (Globecom), Vol. 5, pp.
3173-3177, November 2001, San Antonio (USA).
[113] Q. Liu, E.-H. Yang, Z. Zhang, Throughput analysis of CDMA systems using mul-
tiuser receivers, IEEE Transactions on Communications, Vol. 49 , No. 7, pp. 1192-
1202, July 2001.
[114] P. Kota, C. Schlegel, A wireless packet multiple access method exploiting joint
detection, IEEE International Conference on Communications (ICC), Volume: 4,
pp.2985-2989, 11-15 May 2003.
[115] C. Schlegel, R. Kempter, P. Kota, A novel random wireless packet multiple access
method using CDMA, IEEE Transactions on Wireless Communications, Vol. 5, No.
6, pp. 1362-1370, June 2006.
[116] Karna et al., ”Data transmission method and a radio system, United States Patent
No. 6625138, 23 September 2003.
[117] O. del Rı́o Herrero, R. De Gaudenzi, A high efficiency scheme for quasi-real-time
satellite mobile messaging systems, Proc. of the 10th International Workshop on
Signal Processing for Space Communications, SPSC 2008, pp. 1-9, 6-8 October
2008, Rhodes Island (Greece).
[118] O. del Rı́o Herrero, R. De Gaudenzi, A High Efficiency Scheme for Large-Scale
Satellite Mobile Messaging Systems, 27th AIAA International Communications
Satellite Systems Conference (ICSSC), 1-4 Jun. 2009, Edinburgh (UK).
[119] K. Kim and G. Shevlyakov, Why Gaussianity?, IEEE Signal Processing Magazine,
March 2008, pp. 102-113.
[120] P. Patel, J. Holtzman, Analysis of a DS/CDMA successive interference cancel-
lation scheme using correlations, IEEE Global Telecommunications Conference
(Globecom), Vol. 1, pp. 76-80, 29 November - 2 December 1993, Houston (USA).
[121] C. A. Putman, S. S. Rappaport, D. L. Schilling, A Comparison of Schemes for
Coarse Acquisition of Frequency Hopped Spread Spectrum Signals, IEEE Transac-
tions on Comm., February 1983.](https://image.slidesharecdn.com/resolutionforrandomaccessschemes-220518201640-39d36c53/85/Resolution-for-Random-Access-Schemes-pdf-98-320.jpg)
![BIBLIOGRAPHY 209
[122] G. E. Corazza, P. Salmi, A. Vanelli-Coralli, M. Villanti, Differential and non-
coherent post detection integration techniques for the return link of satellite W-
CDMA systems, 13th IEEE International Symposium on Personal, Indoor and Mo-
bileRadio Communications, Vol. 1, pp. 300- 304, September 2002.
[123] M.K. Sust, R.F. Kaufmann, F. Molitor, G.A. Bjornstrom, Rapid acquisition con-
cept for voice activated CDMA communication, Global Telecommunications Con-
ference (Globecom), Vol. 3, pp. 1820-1826, 2-5 December 1990, San Diego (USA).
[124] M.C. Reed, Acquisition for Satellite UMTS with Large Frequency Offsets, 5th
European Workshop on Mobile/Personal Satcoms (EMPS2002), 25-26 September
2002, Baveno-Stresa, Lake Maggiore (Italy).
[125] 3GPP TS 25.211 v3.12.0, Physical channels and mapping of transport channels
onto physical channels (FDD); Release 1999, September 2002.
[126] 3GPP TS 25.213 v3.9.0, Spreading and modulation (FDD); Release 1999, January
2004.
[127] ETSI TS 101 851-3 V1.1.1, Satellite Component of the UMTS/IMT2000 A-family;
Part 3: Spreading and Modulation December 2000.
[128] ETSI TS 101 851-4 V1.1.1 Satellite Component of UMTS-IMT 2000;A-family;
Part 4: Physical layer procedures, December 2000.
[129] C. E. Gilchriest, Signal-to-Noise Monitoring, JPL Space programs Summary, No.
37-27, Vol. IV, pp.169-184, 30 June 1964.
[130] F. Perez-Fontan, M. Vazquez-Castro, C. Enjamio Cabado, J. Pita Garcia, E. Ku-
bista, Statistical Modelling of the LMS Channel, IEEE Transactions on Vehicular
Technology, Vol. 50, No. 6, November 2001.
[131] F. Perez-Fontan, M. Vazquez-Castro, S. Buonomo, J.P. Poiares-Baptista, B.
Arbesser-Rastburg, S-band LMS channel behaviour for different environments, de-
grees of shadowing and elevation angles, IEEE Transactions On Broadcasting, Vol.
44, No. 1, pp. 40-75, March 1998.
[132] ETSI TS 102 584 v1.1.1 Digital Video Broadcasting (DVB): DVB-SH Implemen-
tation Guidelines, December 2008.
[133] B. Wlodzimierz, The normal distribution: characterizations with applications,
Springer-Verlag 1995, ISBN 0-387-97990-5.
[134] W. Peter, An Introduction to Ergodic Theory, Springer, 1982.
[135] G. Liva, Graph-Based Analysis and Optimization of Contention Resolution Diver-
sity Slotted ALOHA, IEEE Transactions on Communications, Vol. 59, Issue 2, pp.
477-487, February 2011.
[136] E. Paolini, G. Liva, M. Chiani, High Throughput Random Access via Codes on
Graphs: Coded Slotted ALOHA, Proceedings IEEE International Communication
Conference (ICC), 5-9 June 2011, Kyoto (Japan).
[137] H.-C. Bui, J. Lacan and M.-L. Boucheret, An enhanced multiple random access
scheme for satellite communications, In the Proc. of the 2012 Wireless Telecommu-
nications Symposium (WTS), pp. 1-6, 18-20 April 2012, London (United Kingdom).
[138] H.-C. Bui, J. Lacan and M.-L. Boucheret, Multi-slot coded ALOHA with irregular
degree distribution, In the Proc. of the first International IEEE-AESS Conference in
Europe on Space and Satellite Telecommunications (ESTEL), October 2012, Rome
(Italy).](https://image.slidesharecdn.com/resolutionforrandomaccessschemes-220518201640-39d36c53/85/Resolution-for-Random-Access-Schemes-pdf-99-320.jpg)
![210 BIBLIOGRAPHY
[139] F. Ricciato, P. Castiglione, Pseudo-random Aloha for enhanced collision-recovery
in RFID, IEEE Communications Letters, Vol. 17, No. 3, pp. 608-611, March 2013.
[140] P. Castiglione, F. Ricciato, P. Popovski, Pseudo-Random ALOHA for Inter-frame
Soft Combining in RFID Systems, In the Proc. of IEEE 2013 18th International
Conference on Digital Signal Processing. 2013.
[141] J. Massey, P. Mathys, The Collision Channel Without Feedback, IEEE Transac-
tions on Information Theory, Vol. IT-31, No. 2, pp. 192-204, March 1985.
[142] C. Kissling, Performance Enhancements for Asynchronous Random Access Pro-
tocols over Satellite, Proceedings IEEE International Communication Conference
(ICC), pp. 1-6, 5-9 June 2011, Kyoto (Japan).
[143] F. Clazzer and C. Kissling, Enhanced Contention Resolution ALOHA - ECRA,
9th International ITG Conference on Systems, Communications and Coding (SCC),
January 2013, Munich (Germany).
[144] O. del Rı́o Herrero, E. Casini, R. De Gaudenzi, Contention Resolution Diversity
Slotted Aloha Plus Demand Assignment (CRDSA-DA): an Enhanced MAC Protocol
for Satellite Access Packet Networks, 23rd AIAA International Communications
Satellite Systems Conference, 25-28 Sep. 2005, Rome (Italy).
[145] ETSI TS 102 721-5 v1.1.1, Satellite Earth Stations and Systems; Air Interface for
S-band Mobile Interactive Multimedia (S-MIM); Part 5: Protocol Specifications,
Link Layer, December 2011.
[146] Digital Video Broadcasting (DVB), Second Generation DVB Interactive Satellite
System (DVB-RCS2); Guidelines for Implementation and Use of LLS: EN 301
545-2, DVB Document A162, February 2013.
[147] J.O. Irwin, On the Frequency Distribution of the Means of Samples from a Popu-
lation Having any Law of Frequency with Finite Moments, with Special Reference
to Pearson’s Type II, Biometrika, Vol. 19, No. 3/4, pp. 225-239, 1927.
[148] P. Hall, The Distribution of Means for Samples of Size N Drawn from a Population
in which the Variate Takes Values Between 0 and 1, All Such Values Being Equally
Probable, Biometrika, Vol. 19, No. 3/4, pp. 240-245, 1927.
[149] F. Giannetti, M. Luise and R. Reggiannini, Chip timing recovery in digital modems
for continuous-phase CDMA radio communications, IEEE Transactions on Com-
munications, Vol. 43, Issue 234, pp. 762-766, February/March/April 1995.
[150] Newtec, Sat3Play IP Broadband System, http://www.newtec.eu/product/sat3play-
ip-broadband-hub.
[151] Viasat, A Breakthrough in Satellite Communications: The Story of the Technology
Behind ArcLight, Online Journal of Space Communication, Issue 7, VSAT Com-
munication, Fall 2004.
[152] European Space Agency, Advanced MODEM techniques for future satellite access
networks, ESA Contract Number 4000108548/13/NL/JK, 2013.
[153] F. Collard, R. De Gaudenzi, On the Optimum Packet Power Distribution for
Spread Aloha Packet Detectors with Iterative Successive Interference Cancellation,
IEEE Transactions on Wireless Communications, Vol. 13, No. 12, pp. 6783-6794,
October 2014.](https://image.slidesharecdn.com/resolutionforrandomaccessschemes-220518201640-39d36c53/85/Resolution-for-Random-Access-Schemes-pdf-100-320.jpg)
![BIBLIOGRAPHY 211
[154] G. Gallinaro, N. Alagha, R. De Gaudenzi, K. Kansanen, R. Müller and P. Salvo
Rossi, ME-SSA: an Advanced Random Access for the Satellite Return Channel,
IEEE International Conference on Communications (ICC), 8-12 June 2015, London
(UK).
[155] G. Gallinaro, F. Di Cecca, M. Marchitti, R. De Gaudenzi and O. del Rı́o Herrero,
Enhanced spread spectrum ALOHA system level performance assessment, Wiley In-
ternational Journal of Satellite Communications and Networking, on-line published
November 2013.
[156] C. Kissling, A. Munari On the integration of random access and DAMA chan-
nels for the return link of satellite networks, IEEE International Conference on
Communications (ICC), pp. 4282-4287, 9-13 June 2013, Budapest (Hungary).
[157] C. Stefanovic, K.F. Trilingsgaard, N.K. Pratas, P. Popovski, Joint estimation
and contention-resolution protocol for wireless random access, IEEE International
Conference on Communications (ICC), pp. 3382 - 3387, 9-13 June 2013, Budapest
(Hungary).
[158] F. Vazquez-Gallego, M. Rietti, J. Bas, J. Alonso-Zarate, L. Alonso, Performance
Evaluation of Frame Slotted-ALOHA with Succesive Interference Cancellation in
Machine-to-Machine Networks, Proceedings of 20th European Wireless Conference,
pp. 1-6, 14-16 May 2014, Barcelona (Spain).](https://image.slidesharecdn.com/resolutionforrandomaccessschemes-220518201640-39d36c53/85/Resolution-for-Random-Access-Schemes-pdf-101-320.jpg)
Resolution for Random Access Schemes.pdf
- 1. Thesis for the degree of Doctor of Philosophy High Performance Signal Processing-Based Collision Resolution for Random Access Schemes Oscar del Rı́o Herrero Enginyeria i Arquitectura La Salle Universitat Ramon Llull Barcelona 2015
- 2. High Performance Signal Processing-Based Collision Resolution for Random Access Schemes Oscar del Rı́o Herrero European Space Agency The Netherlands E-mail: oscar.del.rio.herrero@esa.int Advisor: Dr. Joan Lluı́s Pijoan Vidal Enginyeria i Arquitectura La Salle, Universitat Ramon Llull, Barcelona, Spain. This thesis has been prepared using L A TEX.
- 3. Abstract Over the past years there has been a fast growing demand for low-cost inter- active satellite terminals supporting both fixed and mobile services, such as consumer broadband access, machine-to-machine communications (M2M), su- pervisory control and data acquisition (SCADA), transaction and safety of life applications. These networks are generally characterized by a large population of terminals sharing the available resources under very dynamic traffic conditions. In particular, in the return link (user to network) of commercial satellite broad- band access networks, residential users are likely to generate a large amount of low duty cycle bursty traffic with extended inactivity periods. A similar situation occurs in satellite mobile networks whereby a large number of terminals typi- cally generate infrequent packets for signalling transmission as well for position reporting or other messaging applications. These services call for the development of efficient multiple access protocols able to cope with the above operating conditions. Random Access (RA) tech- niques are by nature, good candidates for the less predictive, low duty cycle as well as time sensitive return link traffic. Besides, RA techniques are capable of supporting large population of terminals sharing the same capacity and require low terminal complexity. RA schemes have been widely studied and deployed in terrestrial networks, but do not perform well in the satellite environment, which is characterized by very long propagation delays. Today, their use in satellite networks is mainly limited to initial network login, the transmission of control packets, and in some cases, for the transmission of very small volumes of data with very low channel utilization. This thesis proposes three novel RA schemes well suited for the provision of the above-mentioned services over a satellite environment with high performance and low terminal complexity. The new RA schemes are Contention Resolution Diversity Slotted Aloha (CRDSA), Asynchronous Contention Resolution Diver- sity Aloha (ACRDA) and Enhanced Spread Spectrum Aloha (E-SSA), suited for slotted, unslotted and spread spectrum-based systems respectively. They all use strong Forward Error Correction (FEC) codes, able to cope with heavy co-channel interference typically present in RA, and successive interference can- cellation implemented over the successfully decoded packets. The new schemes achieve a normalized throughput above 1 bit/s/Hz for a packet loss ratio below 10−3 which represents a 1000-fold increase compared to Slotted ALOHA. The performance of the proposed RA schemes has been analyzed by means of de- tailed simulations as well as novel analytical frameworks that characterize traffic iii
- 4. iv and packets power statistical distributions, the performance of the FEC coding as well as the iterative interference cancellation processing at the receiver. Keywords: Access control, interference suppression, multiaccess communica- tion, satellite communication, time division multiple access, code division mul- tiaccess, satellite mobile communication, SCADA systems.
- 5. Acknowledgements First of all, I would like to thank my Ph.D. Thesis advisor, Dr. Joan Lluı́s Pi- joan from Escola Universitària d’Enginyeria i Arquitectura La Salle in Barcelona (Spain) for his highly valuable inputs, his academic guidance and for helping me in organizing my research into the form of a thesis. Special thanks to Dr. Riccardo De Gaudenzi, from the European Space Agency in The Netherlands, for his continuous guidance and advice, and for getting as motivated as me in this fascinating field of research. I would also like to thank Enrico Casini and Guray Acar from the European Space Agency in the Netherlands, Gennaro Gallinaro from Space Engineering in Italy, Jean Pierre Choffray from SES-ASTRA in Luxembourg, Daniel Delaruelle from Newtec in Belgium, Antonio Arcidiacono and Daniele Finocchiaro from Eutelsat in France, Pål Orten and Alv Aarskog from Thrane & Thrane in Norway (formerly Nera SatCom). They have all provided highly experienced inputs to the novel random access schemes proposed in this thesis. Norman Abramson from Aloha Networks, with an extensive research in ALOHA based RA schemes, has served as a very important source of inspiration for my PhD. I wish to thank my wife Marta for her strong support and encouragement over the years during which I have studied, my children Claudia and Gabriel for their patience and understanding and my parents for their continuous moral support. This thesis would have not been possible without the support of the European Space Agency, which allowed me to engage myself for many years in this field of research. Finally, I would like to thank Escola Universitària d’Enginyeria i Arquitectura La Salle in Barcelona (Spain) for their kindness and enthusiastic support given to me during all the years I have studied at the University. v
- 7. About the author Oscar del Rio Herrero was born in Barcelona, Spain, in 1971. He received the B.E. degree in Telecommunications and the M.E. degree in Electronics from Universitat Ramon Llull, Barcelona, Spain, in 1992 and 1994, respectively. He received a post-graduate degree in Space Science and Technology with empha- sis in Satellite Communications from the Space Studies Institute of Catalonia (IEEC), Barcelona, Spain, in 1995. He joined ESAs Research and Technology Centre (ESTEC), Noordwijk, The Netherlands, in 1996. In 1996 and 1997 he worked as a Radio-navigation System Engineer in the preparation of the Galileo programme. From 1998 to 2009, he has worked as a Telecommunications Sys- tems Engineer in the Communication - TT&C systems and techniques section. Since 2010 he is working for the Iris programme in the ESA’s Telecommunica- tion Directorate, aiming at the development of a new satellite-based Air-Ground Communication system for Air Traffic Management. In autumn 1994 he started a doctoral programme on Information and Com- munications Technologies and Their Management at Escola Universitària d’En- ginyeria i Arquitectura La Salle, Universitat Ramon Llull, in Barcelona (Spain). During the period 1994 to 1998, he completed his post-graduate Ph.D. courses and initiated himself in the research world. In 2007 he presented his thesis for Diploma d’Estudis Avançats (DEA) with title Signal Processing Based Collision Resolution in Random Access Schemes for Large Satellite Networks. During the period 2001-2014 he has focused his research primarily to multiple access and radio resource management in satellite networks. Research Environment The European Space Agency (ESA) and its 20 Member States work together to pursue a wide range of ambitious and exciting goals in space. Together, they create fascinating projects that would not be feasible for the individual Member States. These projects generate new scientific knowledge and new practical ap- plications in space exploration, and contribute to a vigorous European aerospace industry. ESA has sites in several European countries. The European Space Research and Technology Centre (ESTEC), the largest site and the technical heart of ESA - the incubator of the European space effort - is in Noordwijk, The Netherlands. Most ESA projects are born here, and this is where they are guided through the various phases of development. More than 2000 specialists work here on dozens of vii
- 8. viii space projects. Except for launchers, nearly all ESA projects are managed from ESTEC. In Noordwijk, people work on science missions, on human spaceflight, telecom, satellite navigation, and Earth observation. ESTEC also houses a large pool of people with highly specialized technical knowledge, who are assigned to space projects when their expertise is needed for missions. The Communication - TT&C systems and techniques section is part of the RF Payload Systems division. The group, of about 20 people, is composed of staff, contractors, research fellows, Ph.D. students, graduate trainees and stagiaires. The areas of expertise of the section are: • Telecommunications (physical and medium access control layer techniques, system architectures, advanced digital modems); • Telemetry, tracking and command systems and techniques for spacecrafts targeting deep space, Earth exploration, space operations and telecommu- nication; • Security of TT&C links and Telecom communications. The main roles of the section are to provide support to a wide range of ESA programmes (Telecommunications, Navigation, Deep Space exploration, Earth Observation, Security), ensure the implementation and development of advanced technologies, actively contribute to the development of telecommunication stan- dards in cooperation with European industries and international standardization bodies, provide consultancies to external customers (e.g. EC, SES-ASTRA, Eu- telsat, Inmarsat and others) and perform internal research. The section has performed pioneering work in satellite on-board processing for both regenerative and bent-pipe satellites, advanced CDMA VSAT and mobile networks, multi- user detection and random access for consumer broadband and mobile networks, high-speed modems for high spectral and power efficient modulations, appli- cation of adaptive coding and modulation to Ka-band, satellite UMTS pilot systems and system analysis tools. Summary of contributions This thesis is the result of the research carried out in the following contributions: • Slotted Random Access – E. Casini, R. De Gaudenzi, O. del Rı́o Herrero, Contention Resolu- tion Diversity Slotted Aloha (CRDSA): an Enhanced Random Access Scheme for Satellite Access Packet Networks, IEEE Transactions on Wireless Communications, Vol. 6, Issue 4, pp. 1408-1419, April 2007 – O. del Rı́o Herrero, R. De Gaudenzi, Generalized Analytical Frame- work for the Performance Assessment of Slotted Random Access Pro- tocols, IEEE Transactions on Wireless Communications, Vol. 13, Is- sue 2, pp. 809-821, February 2014
- 9. ix – E. Casini, O. del Rı́o Herrero, R. De Gaudenzi, M. E. Delaruelle, J.P. Choffray, Transmission de donnees par paquets a travers un canal de transmission partage par plusierus utilisateurs, European Patent EP1686746, 30 January 2006 • Unslotted Random Access – O. del Rı́o Herrero, R. De Gaudenzi, High Efficiency Satellite Mul- tiple Access Scheme for Machine-to-Machine Communications, IEEE Transactions on Aerospace and Electronic Systems, Vol. 48, Issue 4, pp. 2961-2989, October 2012 – O. del Rı́o Herrero, R. De Gaudenzi, Methods, apparatuses and system for asynchronous spread-spectrum communication, European Patent EP2159926, 26 August 2008 – R. De Gaudenzi, O. del Rı́o Herrero, G. Acar, E. Garrido Barrabés, Asynchronous Contention Resolution Diversity Aloha: Making CRDSA Truly Asynchronous, IEEE Transactions on Wireless Communica- tions, Vol. 13, Issue 11, pp. 6193-6206, November 2014 – R. De Gaudenzi, O. del Rı́o Herrero, Method and apparatus for trans- mitting data packets over a transmission channel shared by a plurality of users, World Intellectual Property Organization Patent WO/2015/ 000518, 3 July 2013 In addition, the following papers from the author provide complementary infor- mation and analyses to the main area of research of the present thesis: • O. del Rı́o Herrero, R. De Gaudenzi, J. Pijoan Vidal, Design Guidelines for Advanced Random Access Protocols, 30st AIAA International Commu- nications Satellite Systems Conference (ICSSC), 24-27 Sep. 2012, Ottawa (Canada) • O. del Rı́o Herrero, R. De Gaudenzi, A High-Performance MAC Proto- col for Consumer Broadband Satellite Systems, 27th AIAA International Communications Satellite Systems Conference (ICSSC), 1-4 Jun. 2009, Edinburgh (UK) • E. Casini, R. De Gaudenzi, O. del Rı́o Herrero Contention Resolution Diversity Slotted Aloha Plus Demand Assignment (CRDSA-DA): an En- hanced MAC Protocol for Satellite Access Packet Networks, 23rd AIAA International Communications Satellite Systems Conference, 25-28 Sep. 2005, Rome (Italy) • O. del Rı́o Herrero, G. Foti, G. Gallinaro, Spread-Spectrum Techniques for the Provision of Packet Access on the Reverse Link of Next-Generation Broadband Multimedia Satellite Systems, IEEE Journal on Selected Areas in Communications, Vol. 22, Issue 3, pp. 574-583, April 2004 The following publications provide a summary of the new random access schemes proposed by the author:
- 10. x • O. del Rı́o Herrero, R. De Gaudenzi, G. Gallinaro, High-performance ran- dom access schemes, in Co-operative and cognitive satellite systems, S. Chatzinotas, B. Ottersten, R. De Gaudenzi, Ed. Academic Press, 2015, pp. 35-82 • R. De Gaudenzi, O. del Rı́o Herrero, Advances in Random Access Protocols for Satellite Networks, 2009 International Workshop on Satellite and Space Communications (IWSSC 2009), 10-11 September 2009, Siena (Italy)
- 11. Contents I Overview 1 1 Motivation and goals 3 1.1 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 1.1.1 A brief history of satellite communications and services . 3 1.1.2 The multiple access problem . . . . . . . . . . . . . . . . 5 1.1.3 Key terrestrial random access techniques and their appli- cability to satellite . . . . . . . . . . . . . . . . . . . . . . 9 1.2 Goals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 2 The work 19 2.1 Slotted Random Access . . . . . . . . . . . . . . . . . . . . . . . 20 2.2 Asynchronous Random Access . . . . . . . . . . . . . . . . . . . 24 2.2.1 Enhanced Spread Spectrum ALOHA . . . . . . . . . . . . 24 2.2.2 Asynchronous Contention Resolution Diversity ALOHA . 29 3 Outcomes 35 3.1 Prototypes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 3.2 Standards and Systems . . . . . . . . . . . . . . . . . . . . . . . 38 3.3 Example of a M2M satellite network . . . . . . . . . . . . . . . . 42 II Contributions 45 4 Contention Resolution Diversity Slotted Aloha (CRDSA): an Enhanced Random Access Scheme for Satellite Access Packet Networks 47 4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 4.2 System Assumptions . . . . . . . . . . . . . . . . . . . . . . . . . 52 4.3 Proposed Random Access Scheme . . . . . . . . . . . . . . . . . . 52 4.3.1 RA Channel Description . . . . . . . . . . . . . . . . . . . 52 4.3.2 Burst Preamble Issues . . . . . . . . . . . . . . . . . . . . 54 4.3.3 Interference Cancellation Algorithm . . . . . . . . . . . . 56 4.3.4 Analytic CRDSA Throughput Upper Bound Derivation . 57 4.4 Implementation Considerations . . . . . . . . . . . . . . . . . . . 58 4.4.1 CRDSA Modulator Operation . . . . . . . . . . . . . . . 59 4.4.2 CRDSA Demodulator Operation . . . . . . . . . . . . . . 59 xi
- 12. xii CONTENTS 4.5 Numerical Results . . . . . . . . . . . . . . . . . . . . . . . . . . 61 4.5.1 Results Assuming Ideal Channel Estimation for Interfer- ence Cancellation . . . . . . . . . . . . . . . . . . . . . . . 61 4.5.2 Results on Burst Preamble Destructive Collision Probability 67 4.5.3 Results Assuming Non-Ideal Channel Estimation for In- terference Cancellation . . . . . . . . . . . . . . . . . . . . 67 4.6 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70 4.A Probability of Preamble Collision Derivation . . . . . . . . . . . . 72 4.B Probability that i Bursts are Colliding . . . . . . . . . . . . . . . 72 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72 5 Generalized Analytical Framework for the Performance Assess- ment of Slotted Random Access Protocols 75 5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79 5.2 Model without Interference Cancellation . . . . . . . . . . . . . . 80 5.3 Model for S-ALOHA and DSA . . . . . . . . . . . . . . . . . . . 82 5.4 Model with Interference Cancellation . . . . . . . . . . . . . . . . 83 5.4.1 Model for Iterative Interference Cancellation within the Slot of Interest . . . . . . . . . . . . . . . . . . . . . . . . 83 5.4.2 Model for Iterative Interference Cancellation due to suc- cessful detection of replicas . . . . . . . . . . . . . . . . . 87 5.5 Model for CRDSA . . . . . . . . . . . . . . . . . . . . . . . . . . 89 5.6 Summary and Conclusions . . . . . . . . . . . . . . . . . . . . . . 90 5.A FEC Modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92 5.B On the MAI Gaussian Approximation for S-ALOHA . . . . . . . 92 5.C First Order Statistics Sum Lognormal rvs . . . . . . . . . . . . . 95 5.D Assessment probability of loops in CRDSA . . . . . . . . . . . . 96 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97 6 CRDSA patent: Method of packet mode digital communication over a transmission channel shared by a plurality of users 103 7 High Efficiency Satellite Multiple Access Scheme for Machine- to-Machine Communications 107 7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111 7.2 System Requirements and Problem Statement . . . . . . . . . . . 112 7.3 Return Link Multiple Access Messaging Scheme . . . . . . . . . . 113 7.4 Enhanced Spread Spectrum Aloha Detector (E-SSA) . . . . . . . 115 7.4.1 SSA and E-SSA Performance Analysis . . . . . . . . . . . 116 7.4.1.1 SSA Performance Analysis . . . . . . . . . . . . 117 7.4.1.2 E-SSA Analysis . . . . . . . . . . . . . . . . . . 120 7.4.2 Link Margin Optimization for E-SSA . . . . . . . . . . . . 121 7.4.3 Code Collision Probability . . . . . . . . . . . . . . . . . . 122 7.4.4 E-SSA MAC Performance Results . . . . . . . . . . . . . 124 7.4.5 Preamble Receiver Operating Characteristic Performance Assessment . . . . . . . . . . . . . . . . . . . . . . . . . . 130
- 13. CONTENTS xiii 7.5 Uplink RA PHY Layer Adaptations of 3GPP standard . . . . . . 132 7.6 Uplink Packet Transmission Control . . . . . . . . . . . . . . . . 133 7.6.1 Uplink Packet Transmission Control Algorithm . . . . . . 133 7.6.2 SDUPTC Results . . . . . . . . . . . . . . . . . . . . . . . 134 7.6.2.1 System Model . . . . . . . . . . . . . . . . . . . 134 7.6.2.2 Simulation Results . . . . . . . . . . . . . . . . . 136 7.7 Conclusions and Outlook . . . . . . . . . . . . . . . . . . . . . . 138 7.A E-SSA Algorithm Detailed Analytical Description . . . . . . . . . 139 7.B On the Central Limit Theorem Applicability . . . . . . . . . . . 142 7.C Lognormal rv Second Order Moment . . . . . . . . . . . . . . . . 145 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146 8 E-SSA patent: Methods, apparatuses and system for asynchronous spread-spectrum communication 151 9 Asynchronous Contention Resolution Diversity ALOHA: Mak- ing CRDSA Truly Asynchronous 155 9.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159 9.2 ACRDA Concept Description . . . . . . . . . . . . . . . . . . . . 161 9.2.1 ACRDA Modulator . . . . . . . . . . . . . . . . . . . . . 163 9.2.1.1 Baseline . . . . . . . . . . . . . . . . . . . . . . . 163 9.2.1.2 Variant . . . . . . . . . . . . . . . . . . . . . . . 163 9.2.2 ACRDA Demodulator . . . . . . . . . . . . . . . . . . . . 164 9.2.3 ACRDA versus CRDSA . . . . . . . . . . . . . . . . . . . 166 9.3 ACRDA Analytical Performance Derivation . . . . . . . . . . . . 168 9.3.1 Derivation of PLRloop(l) . . . . . . . . . . . . . . . . . . . 172 9.3.2 Derivation of PLRNiter (G, Nrep) . . . . . . . . . . . . . . . 172 9.4 Numerical Results . . . . . . . . . . . . . . . . . . . . . . . . . . 173 9.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182 10 ACRDA patent: Method and apparatus for transmitting data packets over a transmission channel shared by a plurality of users 185 III Concluding remarks 189 11 Conclusions 191 12 Future work 195 Bibliography 198 Acronyms 210
- 17. Chapter 1 Motivation and goals This chapter provides a historical review of satellite communications, in partic- ular from the perspective of the multiple access protocols. From this review, the motivation and goals for the present PhD thesis are derived. 1.1 Motivation 1.1.1 A brief history of satellite communications and services The Space Race (1955-1972) between the Soviet Union (USRR) and the United States (US) opened up the era of satellite communications. In 1957, the Soviet Union launched the first communications satellite (Sputnik 1) into an elliptical Low Earth Orbit (LEO). It was a 58 cm diameter polished metal sphere, with four external radio antennas to broadcast radio pulses. In 1958 the US launched the first satellite which broadcast a recorded Christmas greeting from President Eisenhower during 13 days (Project SCORE). In 1960 the US launched the first active repeater communications satellite (Courier 1B) and was also the first communications satellite powered by solar cells to recharge storage batteries [1]. In 1962 the US launched Telstar 1 and Relay 1 wideband (microwaves) repeater satellites on elliptical LEO orbits, which relayed the first live television signals across the atlantic and pacific oceans respectively. Syncom 3 was the first geostationary (GEO) communication satellite, launched on 1964, and was used to broadcast the 1964 Summer Olympics in Tokyo to the United States. The first commercial geosynchronous satellite, Intelsat I, was launched in 1965, as well as, the first Soviet communication satellite on an inclined Highly Elliptical Orbit (HEO), Molniya 1. Such orbits allowed the satellites to remain visible to sites in polar regions for extended periods, unlike geostationary satellites which appear below the horizon at very high latitudes. In 1974, NASA launched the ATS-6 satellite, which was a precursor to many technologies still in use today on geostationary spacecraft: large deployable antenna (9 m), direct to home (DTH) broadcasting, 3-axis stabilized and antenna pointing through RF sensing. The first in a series of Soviet GEO satellites to carry DTH television was launched on 26 October 1976 (Ekran 1). They were both using ultra-high frequency (UHF) band (around 700-800 MHz), so that the transmissions could be received with 3
- 18. 4 CHAPTER 1. MOTIVATION AND GOALS existing UHF television technology. Due to the limited performance of the first satellites, initial systems had low capacity, required expensive large earth stations with antennae of 15 to 30 m diameter and were used to gather the traffic from an extensive area by means of a ground network and to relay the signals across the oceans (point-to-point trunk connections). During the 1980s and the 1990s, the development of satellite microwave (GHz) technologies, such as contoured multi-beam antennae with higher gain and high power transmission amplifiers (200-300W), enabled the use of smaller earth stations affordable to the end user. These smaller earth stations, typi- cally have a satellite dish antenna ranging from 0.6 to 3.5 m diameter and can be receive-only DTH terminals or bidirectional Very Small Aperture Terminals (VSAT). Their smaller size also enabled mobile services (e.g. on vessels, trains, airplanes). Initial VSAT terminals supported download speeds ranging from few hundred kbps to tens of Mbps and upload speeds ranging from few kbps to few Mbps. The reduced size of satellite terminals resulted in a rapid increase in the number of VSAT networks (including networks with several hundreds of sites), and the emergence of new narrowband and broadband satellite communications services exploiting the satellite’s natural capability to collect or broadcast signals from or to several locations over a very large coverage area: corporate networks, multi-point data transmission networks (retail point-of-sale or automatic teller machine transactions), data collection networks (polling RFID tags, supervisory controlled and data acquisition - SCADA), satellite news gathering, satellite broadcast networks (to cable heads or directly to the consumer) [2]. These satel- lites where typically operating in the C-band (4-8 GHz) and Ku-band (12-18 GHz). Nowadays, most communications GEO satellites carry dozens of transpon- ders, each with a bandwidth of tens of megahertz, providing a very high capac- ity at a reduced communications cost. Communications satellites have a typical launch mass between 3000 and 6000 kg, a payload DC power between 5 kW and 20 kW and a design lifetime of 15 years. Most transponders operate like a ”bent pipe”, relaying back to earth the received signals after performing amplifica- tion and a shift from uplink to downlink frequency. High Throughput Satellites (HTS), such as Anik F2 (2004), Thaicom 4 (2005), Spaceway 3 (2007), KA-SAT (2010), ViaSat-1 (2011), Yahsat Y1B (2012), EchoStar XVII (2012), Inmarsat Global Xpress (2013, 2015, 2016) and Intelsat EpicNG (2016), offer capacities around 100 Gbps which significantly reduces the cost per bit. The majority of HTS operate in Ka-band (20-30 GHz), and the massive increase in capacity is achieved by employing frequency re-use across multiple narrow spot beams (usually in the order of 100s of kilometers per spot beam) [3]. ViaSat -1 satel- lite has 72 spot beams, KA-SAT has 82, EchoStar XVII has 60 and Inmarsat 5 has 89 spot beams. These satellites support two-way communications at higher speeds (e.g. 50 Mbps download and 10 Mbps upload) with small satellite dish antennae (60-70 cm). Typical services offered are broadband connectivity to the consumer and professional markets both fixed and mobile (e.g. land mobile, mar- itime, aeronautical). The Alphasat communications satellite (2013) is another example of a large spacecraft, operating in L-band (1-2 GHz) with a launch mass
- 19. 1.1. MOTIVATION 5 of 6650 kg, a payload DC power of 12 kW and a a 12-meter aperture antenna reflector. L-band satellites are well suited for mobile satellite services requiring low data rates, low cost and small size terminals (e.g. handheld, machine to machine communications). Today, we have around 600 operating communications satellites in orbit, with orders for 20-25 new commercial GEO satellites each year [4]. In 2013 there were about 200 million satellite TV subscribers worldwide (DTH service). These are served primarily via C-band and Ku-band broadcast satellites. To- day they are offering Standard Definition (SD) and High Definition (HD) TV. In the near future they will also serve Ultra HD TV and interactive TV ser- vices. Satellite TV services account for 78% of all satellite services revenues. Regarding the consumer broadband access, there were 2.5 million subscribers (mostly in North America) in 2014 and is expected to grow worldwide to 8.8 million subscribers in 2023, according to Euroconsult forecasts [5]. Enterprise data services are also gradually migrating from broadcast to HTS satellites due to the lower cost per bit, in particular VSAT networks and cellular (3G and 4G) backhaul. Furthermore, it is also foreseen that HTS will also see capac- ity demands coming for commercial mobility (e.g. commercial ships, business jets, commercial airlines). Analysts from Northern Sky Research (NSR) forecast that High Throughput Satellites capacity will increase from 500 Gbps in 2013 to 2 Tbps in 2023 [6]. Another market with a strong growth via satellite are Machine-to-Machine (M2M) communications (e.g. remote control and monitor- ing in transportation and cargo, oil and gas, utilities, military, civil government markets) [7]. Key strengths for M2M via satellite are their reliable connectivity in remote and underserved regions, their global coverage and their independence from terrestrial networks for mission critical applications. The total market fore- cast by NSR in 2023 for satellite M2M services is $ 2.5 billion with 6 million terminals sold. Majority of the market will be served in L-band, being partic- ularly well suited for this type of services with lower cost and smaller physical size terminals. 1.1.2 The multiple access problem In the early years of satellite communications, the services provided were tele- phone and television signals transmission between continents (point-to-point trunk connections). The transmitted signals where analog and each carrier con- veyed either a single TV signal or frequency division multiplexed telephone chan- nels [8]. Multiple access to the satellite was done by fixed Frequency Division Multiple Access (FDMA) assignments to each earth station. For Direct To Home (DTH) television, receive-only terminals are used. The television signals transmitted by the broadcast centre (point-to-multipoint) were initially analog using the NTSC, PAL or SECAM standards and Frequency Mod- ulation (FM). Then, in the mid-1990s digital television standards where adopted exploiting the flexibility to multiplex several TV channels in one carrier and a reduced bandwidth occupation [9], [78]. The digital television signals are trans- mitted by the broadcast centre using a time division multiplex/phase shift keying
- 20. 6 CHAPTER 1. MOTIVATION AND GOALS modulation (TDM/PSK). VSAT networks use bidirectional communications and typically have a star topology where VSATs communicate to a larger master earth station (called hub), but may also communicate directly to other VSATs when in mesh config- uration [11]. The transmitted signals are digital using a PSK modulation. The use of digital transmission offers a greater flexibility to multiplex different types of services (television, voice, data) over the same satellite channel. Digital trans- mission also allows an easier interworking with the digital terrestrial networks. A single outbound carrier provides data transfer from the hub to the remote terminals using a TDM/PSK waveform as for DTH television. The inbound channel (multipoint-to-point) typically consists of several lower bit rate carriers operating in Multi-Frequency Time Division Multiple Access (MF-TDMA). The arrival of the VSAT networks, triggered the need for new multiple access protocols able to efficiently share the satellite communication resources among all VSAT terminals while maintaining acceptable throughout, loss and delay performance. Multiple access is one of the most critical elements to the perfor- mance of VSAT networks. Their design is often constrained by their operational environment. The satellite environment is characterized by radio link impair- ments and a large propagation delay, amongst other inherent attributes (e.g. non-linearities). Typical radio link impairments in the satellite channel are fad- ing and multipath interference. The large propagation delay, in the range of 250 ms for a geostationary satellite, represents a very particular property of the satellite environment that will condition the applicability of terrestrial multi- ple access schemes over the satellite environment. In the satellite channel, the propagation delay is much larger than the time taken to transmit a packet, and a sender may have sent several packets before the receiver starts receiving the first packet. Satellite multiple access schemes must be able to deal with all these characteristics. During the 1980’s and 1990’s numerous multiple access protocols were devel- oped [12, 13, 14, 15, 16, 17, 18, 19]. The available inbound bandwidth resources are divided using FDMA, TDMA, MF-TDMA or Code Division Multiple Access (CDMA) [8]. But as opposed to the fixed assignment multiple access techniques, the new proposed protocols combined demand assignment, random access and fixed assignment to dynamically control the access to the shared resources by the many contending users. These protocols typically have a centralized control (e.g. in the hub) which is managing the access to the satellite resources, as it allows a more efficient multiplexing of services with different priorities (e.g. data, voice, video). First, with the telephone service a portion of the satellite resources (e.g. fre- quency channel in FDMA or timeslots in TDMA) were assigned dynamically on a call basis [20] and the grade of service was determined through the the well-known blocking formulas, such as the Poisson and Erlang B formulas used for terrestrial links [21]. But with the increase of data services (e.g. corporate networks interconnection, retail point-of-sale transactions, SCADA), it was nec- essary to assign satellite resources in a more dynamic way, i.e. on a packet basis rather than call (or circuit) basis. Demand assignment multiple access (DAMA)
- 21. 1.1. MOTIVATION 7 schemes evolved to introduce faster capacity reservation mechanisms in order to reduce end-to-end transmission times. In the CPODA protocol [24] capac- ity reservations can be implemented via contention mini slots and piggybacking them in the header of the scheduled packet transmissions. The CFDAMA-PB protocol [25] behaves like the CPODA protocol, but in addition it randomly assigns the unused traffic slots to inactive users. At low traffic loads the chance that an earth station obtains free-assigned slots is high, thus reducing it’s end- to-end transmission time to a minimum of one satellite hop propagation delay (i.e. 250 ms for GEO satellites). At high loads, the end-to-end transmission de- lay is not reduced as no spare satellite resources are left, but the system remains stable However, the performance of DAMA protocols is highly dependent on the traffic characteristics and the number of earth stations sharing the satellite re- sources. Internet data traffic is known to be statistically self-similar [22, 23]. Satellite terminals generate heavy tailed bursts of packets with a large variance in the inter-arrival times between bursts. Besides, by aggregating streams of such traffic from different earth stations we intensify the self-similarity (bursti- ness) instead of smoothing it (fractal-like behavior). In [26] the performance of the CFDAMA-PB protocol has been analyzed with traffic following a Poisson regime, which is more appropriate for voice, facsimile or SCADA type of services, but not for Internet data traffic as described above. In [27], the author studies the sensitivity of the CFDAMA-PB protocol under different types of Internet data traffic and a large number of earth stations (e.g. larger than the number of traffic slots in one satellite hop propagation delay). It is shown that under bursty traffic and a large number of earth stations the performance of CFDAMA-PB is equivalent to a pure DAMA scheme, i.e. the majority of the free capacity assignments are wasted. Figure 1.1 shows that the throughput component for Free-Assigned slots with bursty traffic is very low at all loads, while for Poisson traffic is quite high at loads up to 80%. Under these conditions, every packet transmission undergoes first a reservation cycle (two satellite hops) and then another satellite hop for the packet transmission (i.e. ≥ 750 ms for a GEO sce- nario) like in a pure DAMA scheme. This poses some limitations on the use of DAMA for consumer broadband access to support real-time or interactive type of applications. To enhance the performance of DAMA protocols for broadband satellite net- works, predictive DAMA protocols have been developed, such as PRDAMA [28]. PRDAMA estimates the positive varying trend of the traffic on each earth sta- tion to allocate the free capacity assignments. Non-linear prediction methods are used to predict the traffic burstiness [29]. This results in a more efficient as- signment of the free bandwidth to those terminals that are predicted to need the resource, thus minimizing the wastage of free bandwidth assignments. However, the effectiveness of the predictive methods cannot be guaranteed for all traffic data [29]. Besides, the improvement in the end-to-end delay with respect to CFDAMA-PB is rather limited (i.e. ∼ 100 ms delay reduction) with an average end-to-end delay over a bent pipe GEO satellite in the order of 750 ms. In [30] and [31] two combined random/reservation multiple access schemes
- 22. 8 CHAPTER 1. MOTIVATION AND GOALS (a) Poisson traffic (b) Bursty traffic Figure 1.1: CFDAMA-PB throughput decomposition in Demand and Free As- signed slots have been proposed dubbed CRRMA and RAN respectively. The basic principle is to perform the first transmission attempt in a random access (RA) slot, thus avoiding the initial two hops satellite delay for capacity reservation, and in case it is unsuccessful (i.e. due to collision), perform a second transmission in a re- served slot. RA performs well in front of bursty traffic, but the improvement in delay of the CRRMA and RAN schemes is limited to low loads, due to the high probability of packet collisions at high loads. Packet collisions are destructive in the considered RA scheme (Slotted Aloha). A review of terrestrial RA tech- niques and their performance over the satellite environment will be the subject of the next section. A final consideration for DAMA schemes is related to the overheads. All DAMA schemes introduce control subframes that are needed by earth stations to make capacity requests and remain synchronized to the satellite network (e.g. for TDMA). This represents an overhead in the inbound channel which can be very large (i.e. comparable to the data subframes) in networks having a very large number of low duty cycle earth stations (e.g. data collection networks with thousands of earth stations such as for SCADA, M2M). In those networks, RA schemes are typically used. The arrival of broadband access and M2M communications has triggered the need for further performance improvements of satellite multiple access pro- tocols. The new multiple access protocols shall provide fast access to a very large population of earth stations (tens of thousands) generating bursty traffic with a low duty cycle, while maintaining acceptable throughout, loss and delay performance. RA protocols represent a good candidate solution, as they are insensitive to the network population size and traffic characteristics, provide low access delays and low complexity on the earth stations. RA protocols used in combination with DAMA are also a good alternative to the free capacity as- signment scheme for the less predictive, low duty cycle as well as time sensitive traffic in broadband access networks.
- 23. 1.1. MOTIVATION 9 1.1.3 Key terrestrial random access techniques and their appli- cability to satellite RA protocols originated in the 1970s from the need for terminal-computer and computer-computer communication. In computer communication, data traffic is bursty. Traffic burstiness is the result of the high degree of randomness seen in the message generation time and size, and of the relatively low-delay constraint required by the user. Users generate traffic with a low duty cycle, but when they do, they require a fast response. As a result, there is a large peak-to-average ratio in the required data transmission rate. In this context, it is not efficient to use fixed channel allocation schemes, as they would result in a low channel uti- lization. A more advantageous approach is to provide a single shared high-speed channel to the large number of users. However, when dealing with shared media conflicts arise when more than one user want to access the shared resources si- multaneously. Therefore, the challenge with shared media is to control the access to the common channel while providing a good level of performance and main- taining a reduced implementation complexity. RA protocols where developed to address this multiple access scenario, and today are extensively used in terres- trial networks over wired and wireless shared media [14, 32, 33]. In this section, we review the key terrestrial RA techniques and analyze their applicability to the satellite environment. One of the most widely used distributed packet access schemes is the Car- rier Sense Multiple Access (CSMA) and its variants. In CSMA, a station senses the medium before transmitting and defers to any ongoing transmission [34]. CSMA/ Collision Detection (CSMA/CD) operates similarly to CSMA, but once the transmission has started, if the sender detects a collision it stops transmitting to reduce the overhead of a collision. When collisions occur, each station willing to transmit backs off for a random time period [35]. Collision detection cannot be implemented over a terrestrial wireless network due to the hidden terminal problem where some stations are out of the transmission and detection range of each other [36]. However it performs well in wired networks and the IEEE has standardized CSMA/CD in the IEEE 802.3 standard [37]. Another variant used in wireless networks is to avoid the collisions similarly to the CSMA/Collision Avoidance (CSMA/CA) scheme. In this scheme the sender tries to avoid a colli- sion after the channel becomes idle, by waiting for an Inter Frame Spacing (IFS) time before contending for the channel. The IEEE has standardized CSMA/CA in the IEEE 802.11 standard [38]. The back-off algorithm in CSMA/CA tries to avoid collisions, but does not remove them all. Small values of the random back- off time cause many collisions while very large values can cause unnecessarily long delays. Secondly, CSMA/CA has failed to solve the hidden terminal prob- lem, and cannot always detect that the medium is busy, thus creating a collision in the channel. All the previous multiple access protocols employ carrier sensing to avoid collisions and offer a good channel utilization, low latency and good stability over channels where packet transmission times are larger than prop- agation delays, but unfortunately cannot operate over satellite channels where propagation delays are very large.
- 24. 10 CHAPTER 1. MOTIVATION AND GOALS Another type of schemes are distributed reservation schemes, also adopted in terrestrial wireless networks. Reservations can be made either through a hand- shaking on the same channel, or through an out-of-band signaling. There is a large variety of hand-shaking protocols. The simplest one is the Multiple Ac- cess Collision Avoidance (MACA) [39] protocol, in which a sender transmits a Request to send (RTS) message to its intended receiver before the data trans- mission. The data is transmitted only after reception of a Clear To Send (CTS) message from the receiver, which the receiver sends on reception of a successful RTS. The MACAW protocol [40] is an extension of the MACA protocol that in- troduces enhancements to the RTS-CTS handshaking by adding a Data Sending (DS) frame by the sender prior to data transmission and an Acknowledgement frame (ACK) by the receiver following the data transmission with the aim of solving the hidden terminal problem. The Floor Acquisition Multiple Access (FAMA) [41] protocol combines carrier sensing with RTS-CTS handshaking, again with the aim of solving the hidden terminal problem. An example of out-of-band signaling protocol is the Busy Tone Multiple Access (BTMA) [36]. In this scheme, a station that is within range and in line- of-sight of all terminals, transmits a busy tone signal on a dedicated busy tone channel, as long as it senses a carrier on the incoming message channel. It is by sensing the busy tone channel that terminals determine the state of the message channel. This scheme solves the hidden terminal problem, but introduces the constraint that a control station needs to be within range of all terminals in the network. Another type of out-band signaling distributed reservation scheme is the Random Access Channel (RACH) used in the third generation (3G) cellular networks [42], where terminals randomly transmit first short packet preambles, and wait for a positive acquisition indicator from the base station prior to the transmission of the complete message (i.e. after successful reservation of the channel). Distributed reservation schemes reduce or eliminate the hidden ter- minal problem typically present in terrestrial radio links, but also rely on short propagation delays, i.e. the reservation delay only represents a small overhead of the total packet transmission time. In the satellite environment, centralized reservation schemes are used instead [25] to avoid several failed attempts prior to the packet transmission, which could represent a very large overhead to the total end-to-end packet transmission delay. The pure ALOHA protocol, first proposed by N. Abramson in [43], is one of the oldest and simplest multiple access protocols. In ALOHA, a terminal trans- mits a packet without checking if any other terminal is active. Within an ap- propriate timeout period, it receives and acknowledgment from the destination, knowing that no conflict has occurred. Otherwise, it assumes that a collision has occurred and must retransmit. To avoid continuously repeated collisions, the re- transmission time is randomized across the terminals, thus spreading the retry packets over time. A slotted version, referred to as Slotted Aloha (S-ALOHA) is obtained by dividing time into slots of duration equal to the duration of a fixed- length packet [44]. Users are required to synchronize the start of transmission of their packets to the slot boundaries. When two packets collide they will overlap completely rather than partially, providing an increase on channel efficiency over
- 25. 1.1. MOTIVATION 11 pure Aloha. The S-ALOHA has the advantage of higher efficiency, but requires time slot synchronization. Both schemes are applicable to the satellite environment, as they have no dependency on the propagation delay. Unfortunately, these schemes are subject to a high collision probability (i.e. no carrier sensing), and their operation in the high load region is not practical in the satellite environment due to the high number of retransmissions required yielding very high latencies. The Diversity Slotted Aloha (DSA) [46] is slightly improving the S-ALOHA performance at low channel loads by sending twice the same packet at random locations in order to increase the time diversity and thus reducing the Packet Loss Ratio (PLR). As for S-ALOHA, operation in the high packet collision probability region is not practical in a satellite environment. It is generally assumed that whenever two packet transmissions overlap in time, these packets destroy each other. This assumption is pessimistic as it neglects capture or near-far effects in radio channels. Capture occurs when a station receives messages simultaneously from two stations, but the signal from one of them drowns out the other, so that no collision occurs. The station with the higher received power is said to have captured the receiver. Some of these effects have been addressed in [44, 45]. Capture is good in the sense that it reduces the time needed in resolving collisions, but may also let weak terminals completely out of the medium. However, in satellite networks the near-far effect is typically very limited. Figs. 1.2 and 1.3 present the performance results for S-ALOHA and DSA in the presence of packets power imbalance following i.i.d. lognormal distributions with equal mean µ and standard deviation σ with both parameters expressed in dB in the logarithmic domain. The results have been obtained by detailed simulations and by using the analytical model derived by the author in [47]. The x-axis represents the normalized average channel MAC load (G) expressed in information bits/symbol, in order to avoid any dependence with the modulation cardinality or coding rate used. As we can see, in both schemes the throughput improves with increasing power imbalance as collisions become easier to resolve (power capture effect). However, as expected, the packet loss ratio is not low and quickly increases as we increase the load on the channel. It can be remarked that for low loads (e.g. G < 0.2), DSA outperforms S-ALOHA. This can be better appreciated in Fig. 1.4, where S-ALOHA and DSA packet loss ratio curves are combined in one figure for the case of no power imbalance and the low load region is expanded. For instance, for a target PLR = 10−2, DSA can achieve a throughput T = 0.05 while for S-ALOHA the maximum achievable throughput is T = 0.01. This is justified by the fact that under light traffic multiple transmission gives better PLR performance. Slotted RA systems require terminals to keep the time slot synchronization. The resulting synchronization overhead greatly reduces the system efficiency, in particular for networks characterized by a large number of terminals with a very low transmission duty cycle like the case in some of the envisaged applications (M2M). Thus, slotted RA is penalizing low-cost terminal solutions. To mitigate this limitation, a pure ALOHA scheme can be employed, but its performance
- 26. 12 CHAPTER 1. MOTIVATION AND GOALS 0 0.5 1 1.5 2 0 0.1 0.2 0.3 0.4 0.5 0.6 Average MAC Channel Load [bits/symbol] Throughput [bits/symbol] Sim. σ=3 dB Sim. σ=2 dB Sim. σ=1 dB Sim. σ=0 dB Ana. σ=3 dB Ana. σ=2 dB Ana. σ=1 dB Ana. σ=0 dB 0 0.5 1 1.5 2 10 −2 10 −1 10 0 Average MAC Channel Load [bits/symbol] Packet Loss Ratio [PLR] Sim. σ=3 dB Sim. σ=2 dB Sim. σ=1 dB Sim. σ=0 dB Ana. σ=3 dB Ana. σ=2 dB Ana. σ=1 dB Ana. σ=0 dB Figure 1.2: Analytical vs simulation S-ALOHA performance for QPSK modu- lation, 3GPP FEC r = 1/2, packet block size 100 bits, Es/N0 = 7 dB in the presence of lognormal packets power imbalance with mean µ = 0 dB, standard deviation σ and Poisson traffic.
- 27. 1.1. MOTIVATION 13 0 0.5 1 1.5 2 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 Average MAC Channel Load [bits/symbol] Throughput [bits/symbol] Sim. σ=3 dB Sim. σ=2 dB Sim. σ=1 dB Sim. σ=0 dB Ana. σ=3 dB Ana. σ=2 dB Ana. σ=1 dB Ana. σ=0 dB 0 0.5 1 1.5 2 10 −2 10 −1 10 0 Average MAC Channel Load [bits/symbol] Packet Loss Ratio [PLR] Sim. σ=3 dB Sim. σ=2 dB Sim. σ=1 dB Sim. σ=0 dB Ana. σ=3 dB Ana. σ=2 dB Ana. σ=1 dB Ana. σ=0 dB Figure 1.3: Analytical vs simulation DSA performance for QPSK modulation, 3GPP FEC r = 1/2, packet block size 100 bits, Es/N0 = 7 dB in the presence of lognormal packets power imbalance with mean µ = 0 dB, standard deviation σ and Poisson traffic.
- 28. 14 CHAPTER 1. MOTIVATION AND GOALS 0 0.05 0.1 0.15 0.2 10 −4 10 −3 10 −2 10 −1 10 0 Average MAC Channel Load [bits/symbol] Packet Loss Ratio [PLR] DSA Sim. σ=0 dB SA Sim. σ=0 dB DSA Ana. σ=0 dB SA Ana. σ=0 dB Figure 1.4: Analytical vs simulation S-ALOHA and DSA performance compar- ison in the low load region for QPSK modulation, 3GPP FEC r = 1/2, packet block size 100 bits, Es/N0 = 7 dB in the presence of no power imbalance and Poisson traffic.
- 29. 1.1. MOTIVATION 15 is worse than for S-ALOHA increasing by two the packet collision probabilities [44]. Direct-sequence spread spectrum (DSSS) multiple access is the most common form of CDMA whereby each user is assigned a particular code sequence which is modulated on the carrier with the digital data modulated on top of that [8]. Users can transmit asynchronously. Even when the same code is used, data can be received [48, 49]. The Spread Spectrum ALOHA (SSA) proposed in [48] has potentially attractive features as it provides a higher throughput capability than S-ALOHA for the same PLR target under equal power multiple access conditions when adopting powerful physical layer FEC (e.g. coding rates ≤ 1/2) and low order modulations (e.g. BPSK, QPSK). In [27] the author shows through simplified analyses that SSA throughput is critically dependent on the demodulator signal-to-noise plus interference (SNIR) threshold. Results reported in [27] indicate that differently from S-ALOHA, SSA shows a steep PLR increase with MAC load. Thus SSA can be operated with low PLR close to the peak of the throughput characteristic. As an example, using turbo codes and relatively small packets, SSA can achieve throughput in the order of T = 0.5 bits/chip for a packet loss ratio of 10−3 (see Fig. 1.5 with σ = 0 dB). This is a very interesting random access scheme for the satellite environment, in particular when asynchronous access is considered. The main reason for the performance improvement of SSA techniques with regards to pure ALOHA is that they can take advantage of a higher traffic aggregation. The average number of packet arrivals over one packet duration can be computed as follows: λ = Nrep G Gp, (1.1) where Nrep is the number of replicas transmitted for each packet, G is the MAC load expressed in information bits/symbol in non-spread systems and bits/chip in spread systems. The processing gain is given by Gp = SF/(r log2 M) where r is the channel coding rate, M is the modulation cardinality and SF is the spreading factor. It can be observed from eqn. (1.1) that large processing gains will increase the value of λ. Typical values for non-spread spectrum systems are λ ≤ 5 while for spread-spectrum systems λ ≈ 100. In Fig. 1.6 we can see that the Poisson PDF normalized to the mean, approaches a delta as we increase λ. This means that the instantaneous number of interfering packets will fluctuate less and less as we increase λ. This is favorable in RA because we can easily make the system to work around a desired load point (i.e. avoid having peaks of traffic) and also because interference can be accurately approximated as AWGN when we have a large number of interferer packets. However, the SSA Achilles’ heel resides in its high sensitivity to multiple access carrier power unbalance. This phenomenon is disrupting the SSA scheme throughput. The SSA throughput is diminished by several orders of magnitude when the received packets power is lognormally distributed with standard devi- ation of 2-3 dB (see Fig. 1.5), as opposed to S-ALOHA and DSA that improved performance with power imbalance due to the power capture effect. SSA is de- signed to work with MAI. Optimal power distribution for SSA is achieved when
- 30. 16 CHAPTER 1. MOTIVATION AND GOALS all packets have the same power level [27, 50]. When there is power unbalance the weakest packets may become not decodable (unless some interference can- cellation technique is applied), thus reducing the RA throughput. DSA and S-ALOHA are designed to work with no or very little interference (collisions are in principle destructive). Under power imbalance some collisions can be recovered (capture effect), thus slightly improving the performance. The above review of known terrestrial RA techniques reveals that none of them is fully performing over the satellite environment. Table 1.1 provides a summary of the different terrestrial RA techniques analyzed in this section. In general, terrestrial RA schemes are known to provide low channel utilization over the satellite environment due to the long propagation delays. Among all of them, the ALOHA-based techniques adapt better to the satellite environment, as they do not have any dependency on the propagation delay, but their performance is constrained by their high collision probability. As a result, today’s satellite sys- tems only use ALOHA-based RA for initial network login, capacity request and short packet transmissions [51, 52]. Therefore, it can be concluded that there is a need to develop new high-performance RA protocols to satisfy the grow- ing market demands for satellite broadband access and M2M communications described in Sects. 1.1.1 and 1.1.2. Technique Main Characteristics Carrier Sense Multiple Access (CSMA) Carrier sense, reduced collision probability, sensitivity to propagation delay. CSMA Collision Detection (CSMA/CD) CSMA with reduced collision overhead. CSMA Collision Avoidance (CSMA/CA) CSMA with collision avoidance mechanism. Multiple Access Collision Avoidance (MACA) Distributed reservation, sensitivity to propagation delay. MACA for Wireless (MACAW) Distributed reservation, sensitivity to propagation delay. Floor Acquisition Multiple Access (FAMA) Distributed reservation and carrier sense, sensitivity to propagation delay. Busy Tone Multiple Access (BTMA) Distributed reservation, sensitivity to propagation delay. 3G Random Access Channel (RACH) Distributed reservation, sensitivity to propagation delay. Slotted ALOHA (S-ALOHA) High collision probability, no dependency propagation delay. Diversity Slotted ALOHA (DSA) S-ALOHA with multiple copies, improved PLR under light traffic. Spread Spectrum ALOHA (SSA) Asynchronous CDMA, no dependency propagation delay, improved PLR under equal packets power. Table 1.1: Summary of Terrestrial Random Access Techniques. 1.2 Goals As described in Sect. 1.1, the improvement of Random Access (RA) protocols performance over satellite could bring great performance advantages to satellite communication networks in serving the growing demand for satellite broadband access and M2M communications. RA protocols provide fast access and are insensitive to the network population size and the traffic characteristics. Unfor- tunately, they offer poor channel utilization over the satellite environment due to the long propagation delays, as well as low reliability due to the high collision probabilities. As we have seen in Sect. 1.1.3, there is a big margin for improve- ment of RA schemes over satellite networks in terms of bandwidth utilization, which is typically not exceeding 0.05 bit/s/Hz if we want to achieve a target Packet Loss Ratio (PLR) < 10−2.
- 31. 1.2. GOALS 17 0 0.2 0.4 0.6 0.8 1 1.2 1.4 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 Average MAC load (bits/chip) Throughput [bits/chip] σ=0dB, semi−ana σ=1dB, semi−ana σ=2dB, semi−ana σ=3dB, semi−ana σ=0dB, simulated σ=1dB, simulated σ=2dB, simulated σ=3dB, simulated 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 10 −8 10 −7 10 −6 10 −5 10 −4 10 −3 10 −2 10 −1 10 0 Average MAC load (bits/chip) Packet Loss Ratio (PLR) σ=0dB, semi−ana σ=1dB, semi−ana σ=2dB, semi−ana σ=3dB, semi−ana σ=0dB, simulated σ=1dB, simulated σ=2dB, simulated σ=3dB, simulated Figure 1.5: Simulated vs. Analytical SSA performance with and without power unbalance from [50]: 3GPP FEC r = 1/3 with block size 100 bits, BPSK mod- ulation, spreading factor 256.
- 32. 18 CHAPTER 1. MOTIVATION AND GOALS 0 2 4 6 8 10 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 x/λ PDF λ=1 λ=2 λ=3 λ=5 λ=6 λ=10 λ=100 Figure 1.6: Traffic probability distribution normalized to the mean value. The main goal of this thesis is to investigate new RA schemes well suited for the satellite environment that can offer high channel efficiency (e.g. > 1 bit/s/Hz aggregate spectral efficiency) and reliability (PLR ≤ 10−3) while keeping very low access delays (in the order of 1 satellite hop delay) and low implementation complexity (comparable to today’s RA solutions). It should be recalled that the satellite beam area is typically very large compared with the terrestrial networks cells, i.e. from several hundreds to a few thousands kilometers of diameter. The new RA protocols shall be robust to operate with very large number of com- municating terminals (e.g. several thousands) sharing the same bandwidth. The new schemes shall also be robust to power imbalance inherent to the satellite communication channel (e.g. due to shadowing and multipath fading). Since we are aiming at low cost, low power and small size earth stations, the new RA schemes require low complexity on earth stations side and shall keep the ad- vanced processing in the receiver side (e.g. in the hub). In addition, both slotted and asynchronous (spread and non-spread) access shall be considered in order to leverage on existing communication standards (e.g. MF-TDMA or CDMA based systems).
- 33. Chapter 2 The work In this thesis, the author presents new advanced random access techniques de- rived from the ALOHA protocols that boost their performance over the satellite environment and enable new market opportunities for satellite communications. As we have seen in Sect. 1.1.3, the RA techniques that best adapt to the satellite environment are the ALOHA-based techniques, as they do not have any depen- dence on propagation delay. This imposes that the new RA schemes shall be able to operate under very high packet collision conditions. To achieve that goal we shall exploit the advances in signal processing tech- niques such as interference cancellation (IC) and strong forward error correction (FEC) coding techniques. The multiuser detection research area covers many different algorithms for mitigating interference between users [53]. While clas- sical IC often assumes synchronous CDMA continuous transmission, this thesis addresses the cases of IC for asynchronous bursty transmissions (spread and non-spread spectrum based), as well as TDMA-based systems. This thesis in- vestigates coordinated IC processing techniques at a common receiver (e.g the hub) with the aim of achieving an aggregated data rate of all simultaneous users approaching the Shannon capacity of the Gaussian noise channel, even for the asynchronous case. The choice of the physical layer FEC scheme for RA systems is not as straight-forward as for conventional communication systems operating in Additive White Gaussian Noise (AWGN). Being a RA Multiple Access (MA) system, the individual but also the aggregate MA performance are relevant. FEC coding schemes shall operate in front of strong interference and not always under AWGN-like conditions (e.g. for TDMA based systems). Therefore lower coding rates, although apparently reducing the individual packet bit rate, may be im- portant to enhance the overall RA scheme performance. This thesis investigates the impact of the FEC coding rate on the RA protocols performance. As ex- plained in Sect. 1.1.3, the level of traffic aggregation in TDMA-based systems is much lower than in CDMA ones. This means that the instantaneous number of interfering packets in TDMA-based systems fluctuates a lot and we cannot guar- antee to operate under a desired working point. This thesis investigates novel approaches to introduce time diversity in TDMA-based systems in order to im- prove the collision probability conditions. The complexity of the new proposed techniques is analyzed to ensure they are commercially viable. 19
- 34. 20 CHAPTER 2. THE WORK This thesis also analyses the performance of the new proposed RA tech- niques and compares them with traditional RA protocols. RA analytical results for conventional RA schemes presented so far were assuming uncoded systems and equi-powered interfering packets [43, 44], which is not what is required in modern satellite communication systems. Some analysis of the power capture effect is reported in [45, 54, 55, 56], but existing models are limited to discrete power levels, do not provide an accurate modeling of the RA Forward Error Cor- rection (FEC) behavior, and more importantly, do not include IC. This thesis defines new RA analytical frameworks that allow calculating the performance of traditional and new RA techniques in terms of PLR and throughput. The proposed frameworks assume arbitrary power and traffic distributions, and ac- curately model the RA FEC and IC behavior. Analytical results have been successfully compared to a wide set of simulation results showing the capability of the proposed analytical framework to predict the the RA scheme performance in the vast majority of application cases. The contributions have been grouped into slotted and asynchronous RA tech- niques and are summarized in the following sections. 2.1 Slotted Random Access Research for slotted RA techniques took place between 2005 and 2009, and then complemented by an analytical framework in 2014. The research on slotted sys- tems was triggered by the need to introduce efficient and reliable RA schemes to the existing satellite systems using standards for interactive satellite broadband networks like the Digital Video Broadcasting (DVB) Return Channel via Satel- lite (DVB-RCS) [51] and the Telecommunication Industry Association (TIA) IP over Satellite (IPoS) [52]. The standard-based and many proprietary satellite communication systems employ MF-TDMA access and rely primarily on DAMA schemes for the transmission of data traffic on the return link (earth station to hub), as described in Sect. 1.1.2. The arrival of broadband Internet access for the consumer market introduced the need for a fast satellite access to a very large population of earth stations sharing the same bandwidth. The first contribution from the author is a RA scheme dubbed Contention Resolution Diversity Slotted Aloha (CRDSA) [57] and included in this thesis in chapter 4. The CRDSA technique consists in transmitting every burst two or more times (Nrep) in randomly selected slots within a TDMA frame of Nslots. Each burst contains in the header information of the location of the replicas within the frame (see Fig. 2.1). The transmitter achieves network synchro- nization in the same way as it is defined in the MF-TDMA based standards [51, 52], i.e. the terminal exploits the forward link network clock reference re- quired for MF-TDMA terminal time reference generation to also generate time and frequency reference for CRDSA. The complete TDMA frame is sampled and stored in a digital memory at the receiver side. By using simple yet efficient interference cancellation tech- niques, clean bursts are recovered (e.g. packet 3 in slot 5 in Fig. 2.1) and the interference generated by their replicas on other slots are cancelled (e.g. packet
- 35. 2.1. SLOTTED RANDOM ACCESS 21 2 2 1 3 3 4 4 6 5 6 5 1 TDMA RA frame with Nslots TDMA RA frame with Nslots RA slot RA slot Figure 2.1: Example of CRDSA frame with two replicas per packet. 3 in slot 4). By performing an iterative processing of the TDMA frame it is proven that most of the initial collisions can be resolved. In Fig. 2.2 it is shown that CRDSA largely outperforms classical Slotted Aloha techniques in terms of throughput and packet loss ratio (PLR). For a MAC packet loss probability of 2%, channel utilization efficiencies close to 40% can be achieved with the CRDSA technique (see Fig. 2.2(b)), while for S-ALOHA only a channel utilization effi- ciency of 2% can be achieved for the same packet loss probability (see Fig. 1.4). This represents a 20-fold throughput improvement compared to S-ALOHA. It is important to highlight the importance of PLR improvement in the satellite environment, as it has a direct benefit on the end-to-end delay due to a reduction in the need for packet retransmissions. The author’s contribution [57] presents detailed performance results on delays. The performance boost of the CRDSA protocol is achieved thanks to the im- plementation on the hub receiver of a collision resolution capability that exploits IC techniques. IC techniques have been largely investigated for Code Division Multiple Access (CDMA) [53] but, at the author’s knowledge, have never been proposed in a TDMA S-ALOHA context. CRDSA is basically a TDMA access scheme that uses the information from the successfully decoded packets to cancel the interference their replicas may generate on other slots. We are able to cancel the interference because we know the data symbols of the interference from the successfully decoded packets. One of the main issues in applying IC techniques to TDMA S-ALOHA is related to the need for accurate channel estimation for the burst replicas removal where collision(s) occur. In fact, collisions in a satel- lite TDMA multiple access channel are typically destructive as the near-far effect is rather limited. While carrier frequency, amplitude and timing estimation for IC can be derived from the ”clean” replica (e.g. packet 3 in slot 5 in the example from Fig. 2.1), carrier phase has to be estimated on the slot where collision(s) occurs (e.g. in slot 4). This is because the phase in practical broadband systems is time variant also from slot to slot. This key problem has been solved in [57] by exploiting the burst preamble which is individually ”signed” by a pseudo- random binary sequence randomly selected among the available code family by each active earth station for each burst in each frame. All replicas of the same burst use the same preamble code. This approach does not require a centralized preamble code assignment, thus allowing to maintain the random access nature of the proposed scheme. In this way the preamble can be used for carrier phase
- 36. 22 CHAPTER 2. THE WORK 0 0.2 0.4 0.6 0.8 1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 Normalized Load (packets/slot) Throughput SA DSA CRDSA (a) Throughput versus load 0 0.2 0.4 0.6 0.8 1 10 −6 10 −5 10 −4 10 −3 10 −2 10 −1 10 0 Normalized Load (packets/slot) Packet Loss Ratio SA DSA CRDSA (b) Packet Loss Ratio versus load Figure 2.2: Simulation S-ALOHA, DSA and CRDSA performance for Nrep = 2, Nslots = 100, QPSK modulation, 3GPP FEC ρ = 1/2, packet block size 100 bits, Es/N0 = 7 dB in the presence of no power imbalance and Poisson traffic.
- 37. 2.1. SLOTTED RANDOM ACCESS 23 estimation also in case of multiple collisions which normally are destructive for channel estimation and payload decoding. With the above channel estimation implementation, it is shown in Section V.C in from [57] that CRDSA perfor- mance results on non-ideal channel estimation for IC are very close to those derived with ideal channel estimation. An alternative way to perform channel estimation is to use a single Unique Word (UW) for all bursts and all earth stations, but use some of the payload symbols, which are known from the clean replicas, for carrier phase estimation. Payload symbols are uncorrelated among the colliding packets, even though with- out excellent cross-correlation properties. It is found that by using payload sym- bols for carrier phase estimation, performance results are comparable to those obtained by using multiple pseudo-random binary sequences for the burst pream- bles. But his approach reduces complexity on both the earth stations and hub as a single UW is used. Besides, the knowledge of the payload data allows to track the carrier phase along the burst by using an normal Data Aided-Phase Locked Loop (DA-PLL), which is particularly useful when low cost terminals with very low baud rate are used (e.g for interactive TV), since the RF front ends introduce a phase noise (nuisance) making the carrier phase not constant over the burst duration. This improvement was proposed by Daniel Delaru- elle from NEWTEC and Jean Pierre Choffray from SES-ASTRA. Both channel estimation approaches have been extensively investigated in [58] (see chapter 6). Prior to the IC of a packet, the whole packet is re-modulated allowing data aided refined amplitude and phase estimation. This maximizes the IC efficiency with a minimal degradation of the CRDSA performance when compared to ideal channel estimation conditions [58]. CRDSA has a number of parameters that can greatly influence their per- formance, being the most important ones FEC coding rate ρ, the number of replicas Nrep and the packets power imbalance. With the aim of optimizing its performance, the author has developed an analytical framework and a de- tailed simulator for the performance assessment of slotted RA protocols (see [47] included in this thesis in chapter 5). The proposed analytical framework assumes arbitrary power and traffic distributions, and accurately models the FEC and RA IC behavior. To be remarked that using a powerful FEC code (e.g. ρ = 1/3) and an adequate signal to noise ratio (e.g. SNR = 10 dB), there is non zero probability of correctly detecting the packet even in presence of a collision (capture effects). Besides FEC coding schemes shall operate in front of strong interference and not always under AWGN-like conditions (e.g. for TDMA based systems). When payload burst are protected by a powerful FEC, a phase estimation error standard deviation up to 15 degrees can be tolerated without significant impact on the performance [58]. Therefore lower coding rates, al- though apparently reducing the individual packet bit rate, are very important to enhance the overall RA scheme performance. Three burst replicas gives bet- ter performance than two, because it significantly reduces the loop occurrence probability between packets from different earth stations. In the example if Fig. 2.1, packets 4 and 5 have formed a loop as their replicas have been transmitted in the same slots. The issue of loops has been extensively analyzed by the author
- 38. 24 CHAPTER 2. THE WORK in [47], and with three packet replicas the probability of occurrence is well below the packet loss ratio target for our applications (PLR < 10−3). Packets power imbalance greatly contributes to the interference cancellation process and boosts the performance of CRDSA. Strong packets are decoded first and weaker packets are successively decoded following the iterative IC process. Power imbalance re- sults from the combined effect of the time variant atmospheric propagation, open loop power control errors (if applicable), earth station equivalent isotropically radiated power (EIRP) and satellite receive antenna gain variations. It is shown in Fig. 2.3 that CRDSA with FEC ρ = 1/3, Nrep = 3 and power imbalance following a lognormal power distribution with standard deviation σ = 3 dB, can achieve a throughput T = 1.4 bits/symbol for a PLR < 10−3, which represents a 1400-fold improvement with respect to S-ALOHA. 2.2 Asynchronous Random Access Research on asynchronous random access took place between 2008 and 2014. The research on asynchronous access was triggered by the growing demand for low cost and highly efficient solutions supporting both fixed and mobile satellite services, such as consumer broadband access, machine-to-machine communica- tions (M2M), interactive TV, supervisory control and data acquisition (SCADA), transaction and safety of life applications. These networks are generally char- acterized by a large population of low power and small size terminals, sharing the available resources under very dynamic traffic conditions. In particular, in the return link (earth station to hub) users are likely to generate low duty cy- cle bursty traffic with extended inactivity periods. Slotted RA systems require terminals to keep the time slot synchronization through a control loop. The resulting synchronization overhead can greatly reduce the system efficiency, in particular for networks characterized by a very large number of terminals with a very low transmission duty cycle, like it is the case in some of the envisaged appli- cations (e.g. M2M). Thus, slotted RA is penalizing low-cost terminal solutions. To mitigate this limitation, asynchronous RA schemes are preferable. Asynchronous RA schemes can been divided into spread spectrum and non- spread spectrum multiple access techniques. This thesis presents two novel asynchronous RA schemes. The first one, dubbed Enhanced Spread Spectrum ALOHA (E-SSA), is an extension of Spread Spectrum ALOHA presented in Section 1.1.3 and introduces an IC mechanism at the receiver. The second RA technique, Asynchronous Contention Resolution Diversity ALOHA (ACRDA), is an extension of the CRDSA protocol presented in Sect. 2.1 to a truly asyn- chronous multiple access environment. ACRDA is non-spread spectrum based, but incorporates some of the E-SSA receiver processing functions to support the asynchronous access from all the earth stations to the shared medium. 2.2.1 Enhanced Spread Spectrum ALOHA Enhanced Spread Spectrum ALOHA (E-SSA) behaves like SSA on the trans- mitter side (see Sect. 1.1.3). The E-SSA novelty lies on the detector [59]. The
- 39. 2.2. ASYNCHRONOUS RANDOM ACCESS 25 0 0.5 1 1.5 2 0 0.5 1 1.5 Average MAC Channel Load − G [b/s/Hz] Normalized Throughput [T] Sim. σ=3 dB Sim. σ=2 dB Sim. σ=1 dB Sim. σ=0 dB Ana. σ=3 dB Ana. σ=2 dB Ana. σ=1 dB Ana. σ=0 dB (a) Throughput 0 0.5 1 1.5 2 10 −7 10 −6 10 −5 10 −4 10 −3 10 −2 10 −1 10 0 Average MAC Channel Load − G [b/s/Hz] Packet Loss Ratio [PLR] Sim. σ=3 dB Sim. σ=2 dB Sim. σ=1 dB Sim. σ=0 dB Ana. σ=3 dB Ana. σ=2 dB Ana. σ=1 dB Ana. σ=0 dB (b) PLR Figure 2.3: Analytical vs. Simulation CRDSA performance for Nrep = 3, Niter max = 15, Nslots=1000, QPSK modulation, 3GPP FEC ρ = 1/3, packet block size 100 bits, Es/N0 = 10 dB in the presence of lognormal packets power unbalance with mean µ = 0 dB, standard deviation σ and Poisson traffic [47].
- 40. 26 CHAPTER 2. THE WORK 4000 4050 4100 4150 4200 4250 4300 0 100 200 300 400 500 600 Time Number of interfering packets Packet duration=312; Normalized Load=0.5 Figure 2.4: Realization of the instantaneous and average number of interfering packets over packet of interest [60]. detector located at the hub is the heart of the system, as it has to support a high throughput of incoming packets. For a typical spread spectrum system with Nrep = 1, SF = 256, r = 1/3, M = 2 and a normalized load G = 0.5, the aver- age number of packet arrivals over a packet duration is λ = 384 (see eqn. (1.1)), which represents quite a large value. In Fig. 2.4, the author shows that the instantaneous number of interfering packets is almost constant around the mean value λ = 384 (red line) over the whole packet duration. This is advantageous for a RA scheme design because the interference can be accurately approximated as AWGN when we have a large number of interferer packets. The time on the x-axis corresponds to symbol time steps, and the packet shown in this example is a 100 bit packet with 300 coded bits. The complete set of system parameters used for the E-SSA analyses can be found in Table I from [50], also included in this thesis in chapter 7. The principle of the E-SSA detector is illustrated in Fig. 2.5. The received CDMA signal is band-pass filtered, sampled, digitally down-converted to base- band with I-Q components separation and stored in a digital memory of 2·WNc s real samples. We assume that complex baseband samples are stored in I-Q for- mat. Nc s corresponds to the number of chips per physical layer channel symbol and W corresponds to the memory window size in symbols for the interference cancellation process at the hub. The window size W shall be optimized to be the smallest possible value yielding good IC performance. Typically, W should
- 41. 2.2. ASYNCHRONOUS RANDOM ACCESS 27 1 5 2 3 4 6 8 7 10 9 12 15 13 14 11 17 18 19 16 Sliding Window Time Iterative IC process within window W+(n-1) W W+n W W Figure 2.5: Enhanced Spread Spectrum Aloha algorithm description. be three times the physical layer packet length in symbols. The receiver window memory is shifted in time in discrete steps allowing some overlap of packets on each window step (sliding window process). To reduce the detector complexity the window step ∆W, in symbols, shall be the largest possible yielding a good IC performance. Normally, the window step ∆W shall be between 1/3 to 1/2 of the window length W depending on its size. At each window step, the following iterative Successive Interference Cancellation (iSIC) process takes place: 1. Store in the detector memory the new baseband signal samples correspond- ing to the current window step (n); 2. Perform packets preamble detection and select the packet with highest Signal to Noise plus Interference Ratio (SNIR) value; 3. Perform data-aided channel estimation for the selected packet over the preamble; 4. Perform FEC decoding of the selected packet; 5. If the decoded FEC frame is considered correct after CRC check then: (a) Perform enhanced data-aided channel estimation over the whole re- covered packet (carrier frequency, phase, amplitude, timing); (b) Reconstruct at baseband the detected packet for following cancella- tion step; (c) Perform interference cancellation; 6. Repeat from step 2 until NW it max iterations are performed. When the limit is reached, advance the observation window by ∆W.
- 42. 28 CHAPTER 2. THE WORK A detailed analytical description of the E-SSA algorithm is provided in Ap- pendix I from [50], also included in chapter 7 of this thesis. Two important steps for an optimal E-SSA performance are the packet preamble detection in front of very high interference (step 2) and the IC process defined in step 5(c). The hub demodulator starts searching for the presence packet preamble by means of a conventional preamble correlator. Because of the incoming packets car- rier frequency uncertainty due to the oscillator instabilities, preamble parallel search in the frequency domain is typically required. The preamble acquisition performance needs to be lower than the target PLR (e.g 10−3) in terms of the probability of missed detection Pmd and the probability of false alarm Pfa at the very high MAC operating load allowed by E-SSA. Preamble acquisition aspects have been summarized in Sect. IV.E in [50]. A particular aspect in SS is the need for a longer preamble for packet detection due to operation under very high MAI (Ec/(N0 + I0) ≈ −30 dB). When a preamble is detected, the associ- ated packet amplitude, carrier frequency, chip timing and phase are estimated exploiting the preamble plus possible auxiliary pilot symbols. At this point, de- spreading, coherent detection, payload and signalling bits decoding takes place. This demodulation step is comparable to a conventional DS-SS burst demodu- lator. The transmitted packet contains also a CRC, which is exploited by the demodulator to verify that the packet has been correctly detected. If this is the case, the payload information is extracted. At the same time the decoded bits are re-encoded and re-modulated. This locally regenerated packet is then correlated with the memory samples corresponding to the currently detected packet loca- tion. Through this correlation process it is possible to extract a more accurate amplitude and carrier phase estimation compared to the initial demodulation step. This is because the replica packet(s) amplitude and phase estimation is now extending over the whole packet duration instead of being limited to the sole preamble length. In this way, amplitude and phase variations over the packet duration can be estimated resulting in a more accurate cancellation process. This key step is called payload data-aided refined channel estimation and is the same used for CRDSA IC [58]. Thanks to the refined channel estimation, the E-SSA demodulator achieves accurate cancellation of the detected packet from the sliding window memory. The E-SSA performance has been investigated in-depth both by analysis and simulation in Sect. IV.D from [50]. The analytical framework proposed by the author, accurately models the performance of spread-spectrum RA techniques, such as SSA and E-SSA, and assumes arbitrary power and traffic distributions, and accurately models the FEC and RA IC behavior (see Section IV.A from [50]). Fig. 2.6 reports some key conclusion about E- SSA performance. Performance results shall be compared to the SSA ones reported in Fig. 1.5. First of all, assuming a target PLR of 10−3 and no power imbalance, the E-SSA throughput is 1.12 bits/symbol, i.e. 2.4 times higher than conventional SSA. When lognor- mal distributed packet power is assumed (with standard deviation σ = 2 dB), then the E-SSA throughput reaches 1.7 bits/symbol, i.e. 17 times larger than SSA. This striking result is due to the iSIC superior performance compared to a conventional SSA burst demodulator in case of imbalanced power. The received
- 43. 2.2. ASYNCHRONOUS RANDOM ACCESS 29 packets power fluctuation at the hub is caused by the uplink Land Mobile Satel- lite (LMS) channel shadowing and multipath fading, as well as the satellite and earth station antenna gain variations as a function of the location. In Sect. VI from [50], Riccardo De Gaudenzi proposes a SNIR driven uplink packet transmis- sion control (SDUPTC). SDUPTC is based on the principle to transmit packets only when the downlink signal quality is good enough, i.e. the signal strength or better SNR is within a certain window representative of Line-of-Sight (LOS) conditions. If this is not the case the transmission is delayed until (quasi)-LOS conditions are verified. The proposed simple SDUPTC algorithm is intended to maximize the probability of packet successful transmission, limit the power imbalance among the packets received at the hub side to improve the RA scheme performance and to avoid creating multiple access interference (MAI) when not useful for packet transmission. 2.2.2 Asynchronous Contention Resolution Diversity ALOHA The Asynchronous Contention Resolution Diversity ALOHA (ACRDA) proto- col is analyzed in detail in [61], and included in this thesis in chapter 9. An introduction to the ACRDA patent [62] is also presented in chapter 10. Unlike S-ALOHA, DSA or CRDSA, ACRDA eliminates the need to maintain accurate slot synchronization among all transmitters, but differently to SSA and E-SSA, does not require the use of spread spectrum techniques. The need for transmit- ter synchronization is a major drawback for very large networks (e.g. M2M), as the signaling overhead scales up with the number of transmitters independently from their traffic activity factor. In ACRDA, earth stations behave like in CRDSA, exploiting packet replicas and the associated location signaling, but do not need to maintain slot synchro- nization with the hub. The term Virtual Frame (VF) is introduced here to refer to the concept of frame of slots that is only local to each transmitter. In ACRDA, for all transmitters, each VF is composed of a number of slots Nslots and each slot has a duration Tslot with an overall frame duration Tframe = Nslots · Tslot. Fig. 2.7 shows the Virtual Frame compositions for the ACRDA scheme. The different transmitters are not time synchronized, and hence, the time offset be- tween VF(i), VF(i − 1), VF(i + 1) is arbitrary. In case of mobile applications, the Doppler effect may have an appreciable impact in terms of incoming packets clock frequency offset. In this situation the VFs will have slightly different dura- tion. However, the localization process of the replica packets within each VF will remain accurate, as the hub burst demodulator will extract for each VF its own clock reference. Since the demodulator has no knowledge of the start of VFs, the signaling of the replicas location within the VF has to be relative to each packet position and not absolute to the start of the VF. Two options exist for the ACRDA modulator (as described in Sect. II.A from [61]), which consist in transmitting all packet replicas randomized within the VF, or transmitting the first packet replica in the first slot of the VF and the remaining replicas random- ized within the VF. The latter option, improves end-to-end delay performance since first packet replica is transmitted without any added delay.
- 44. 30 CHAPTER 2. THE WORK 0 0.5 1 1.5 2 2.5 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 MAC load (bit/s/Hz) Throughput (bit/s/Hz) σ=0 dB, ana. σ=1 dB, ana. σ=2 dB, ana. σ=3 dB, ana. σ=0 dB, sim. σ=1 dB, sim. σ=2 dB, sim. σ=3 dB, sim. (a) E-SSA Throughput with and without power unbalance. 1 1.5 2 2.5 10 −8 10 −7 10 −6 10 −5 10 −4 10 −3 10 −2 10 −1 10 0 MAC load (bit/s/Hz) Packet Loss Ratio (PLR) σ=0 dB, ana. σ=1 dB, ana. σ=2 dB, ana. σ=3 dB, ana. σ=0 dB, sim. σ=1 dB, sim. σ=2 dB, sim. σ=3 dB, sim. (b) E-SSA PLR with and without power unbalance. Figure 2.6: Simulated vs. Analytical E-SSA performance with and without power unbalance, 3GPP FEC r = 1/3 with block size 100 bits, BPSK modulation, spreading factor 256 [50].
- 45. 2.2. ASYNCHRONOUS RANDOM ACCESS 31 4 4 4 Virtual Frame (i-1) 5 5 5 Virtual Frame (i) 3 3 3 Virtual Frame (i-2) 6 6 6 Virtual Frame (i+1) 7 7 7 Virtual Frame (i+2) 2 2 2 Virtual Frame (i-3) 1 1 1 Virtual Frame (i-4) 8 Virtual Frame (i+3) 9 Virtual Frame (i+4) 9 9 8 8 ∆W ∆W W(s) W(s) W(s + 1) W(s + 1) Offset between VF(i-1) and VF(i) Offset between VF(i) and VF(i+1) Figure 2.7: Virtual frame compositions in ACRDA. ACRDA combines features of CRDSA, and E-SSA on the receiver side. On one hand, the same E-SSA sliding window-based memory processing described in Sect. 2.2.1 is adopted to handle the asynchronously arriving packet replicas. On the other hand, for a given receiver memory position, the replica packets cancellation scheme described in Sect. 2.1 is borrowed from the CRDSA de- modulator processing. The received signal is sampled at baseband, and complex signal samples are stored in a sliding window memory of W virtual frames (see Fig. 2.7). Typically W = 3 for optimal ACRDA performance, as for E-SSA. For a given window position, the demodulator performs the same iterative pro- cessing as for CRDSA to decode the clean packets and perform IC of the replica packets. The detailed ACRDA demodulator operation is described in Sect. II.B from [61]. Once the iterative processing of a specific window position is com- pleted, the sliding window is shifted towards the right in time by ∆W, which is a fraction of the VF duration (e.g. ∆W = 1/6). The oldest samples spanning the leftmost part of the memory will be removed and the emptied rightmost part of the memory will be then filled with the new incoming complex samples. ACRDA provides better PLR and throughput performance than CRDSA. Fig. 2.8 compares the performance of the ACRDA vs. CRDSA, evaluated via mathematical analysis and computer simulations. Similarly to the other pro- posed RA schemes, the author proposes in Sect. III from [61] an analytical framework derived from the RA analytical framework presented in [47]. For a target PLR = 10−3, ACRDA channel can be loaded up to 1 bits/symbol, while CRDSA not more than 0.75 bits/symbol. It is important to note that the PLR floor present at low loads (i.e. G < 1 bits/symbol) is lower for ACRDA than CRDSA. The reason for this is that the probability of loops occurrence,
- 46. 32 CHAPTER 2. THE WORK described in Sect. 2.1, is significantly lower in ACRDA than in CRDSA due to the asynchronous nature of the access. The effect of loops in performance is an- alyzed in detail in Sect. III from [61] for ACRDA and in Appendix D from [47] for CRDSA. To mitigate this limitation, the optimal configuration for CRDSA is with Nrep = 3 as presented in Fig. 2.3. In ACRDA, slightly better performance can be already achieved with Nrep = 2, thus reducing the demodulator com- plexity. It shall be recalled that ACRDA also reduces earth station complexity, as it does not require any control loop to keep time slot synchronization with the hub. Similarly to CRDSA, ACRDA performance improves in front of power imbalance. Fig. 8 from [61] shows that ACRDA with FEC ρ = 1/3, Nrep = 2 and power imbalance following a lognormal power distribution with standard deviation σ = 3 dB, can achieve a throughput T = 1.5 bits/symbol for a target PLR < 10−3. In Sect. IV from [61], Guray Acar and Eloi Garrido analyze and compare the delay performance between ACRDA and CRDSA. Results show that ACRDA reduces the delay by a factor of 10 compared to CRDSA for low loads (e.g. G = 0.3 bits/symbol) and a factor of 2 at high loads (e.g. G = 0.9 bits/symbol). In Fig. 2.9, the performance of ACRDA is compared against E-SSA. As we can see, in particular for the case of large power imbalance, E-SSA performance is better than ACRDA. The use of QPSK in ACRDA improves the efficiency, but also increases the sensitivity to channel estimation errors. Besides, as we have seen in Fig. 1.6, spread-spectrum achieves a higher aggregation of traffic with lower interference fluctuations which is favorable for RA.
- 47. 2.2. ASYNCHRONOUS RANDOM ACCESS 33 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 0 0.2 0.4 0.6 0.8 1 Average MAC Channel Load − G [bits/symbol] Throughput − T [bits/symbol] ACRDA Simulation ACRDA Analytical CRDSA Analytical CRDSA Simulation (a) Throughput 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 10 −6 10 −5 10 −4 10 −3 10 −2 10 −1 10 0 Average MAC Channel Load − G [bits/symbol] Packet Loss Ratio [PLR] CRDSA Simulated ACRDA Simulated ACRDA Analytical CRDSA Analytical (b) PLR Figure 2.8: Simulation and analytical ACRDA and CRDSA performance for Nrep = 2, Nslots=100 (simulations), QPSK modulation, 3GPP FEC ρ = 1/3, packet block size 100 bits, Es/N0 = 10 dB in the presence of no packets power imbalance and Poisson traffic, window size W = 3 virtual frames and a window step ∆W = 0.15 [61].
- 48. 34 CHAPTER 2. THE WORK 0 0.5 1 1.5 2 2.5 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 Average MAC Channel Load − G [bits/s/Hz] Throughput − T [bits/s/Hz] E−SSA σ=0 dB E−SSA σ=3 dB ACRDA σ=0 dB ACRDA σ=3 dB (a) Throughput 1 1.2 1.4 1.6 1.8 2 10 −6 10 −5 10 −4 10 −3 10 −2 10 −1 10 0 Average MAC Channel Load − G [bits/s/Hz] Packet Loss Ratio (PLR) E−SSA σ=0 dB E−SSA σ=3 dB ACRDA σ=0 dB ACRDA σ=3 dB (b) PLR Figure 2.9: Simulated E-SSA performance with BPSK modulation, 3GPP FEC ρ=1/3, packet block size 100 bits, spreading factor 256 vs. ACRDA performance for Nrep = 2, Nslots=100, QPSK modulation, 3GPP FEC ρ=1/3, packet block size 100 bits, Es/N0 = 10 dB, window size W = 3 virtual frames and a window step ∆W = 0.15, with and without power imbalance and Poisson traffic.
- 49. Chapter 3 Outcomes The contributions from the author in this PhD thesis have derived in a number of prototyping activities, commercial products and telecommunications standards. They are briefly presented in this chapter. 3.1 Prototypes A real-time prototype of a CRDSA modulator and demodulator has been de- veloped by Space Engineering (Italy) in 2011 in the frame of an ESA contract [63, 64]. Fig. 3.1 compares the measured results in the test bed against those obtained by simulations. We can see that experimental results are very close to the simulation findings assuming an ideal CRDSA demodulator. A reference implementation of the E-SSA technique based on a software defined radio (SDR) platform was developed by MBI (Italy) in 2012 under Eu- telsat’s specifications [65, 66, 67]. The MBI prototype allowed to fully verify the E-SSA performance in a real-time environment including preamble detec- tion, channel estimation, interference cancellation and frequency errors. We can see in Fig. 3.2 that the measured performance are very close to the simulated ones. The PLR remains very low until almost the peak MAC throughput and, as expected, power imbalance improves the E-SSA performance. A second E-SSA prototype was developed in 2014 in the frame of an ESA project, called ANTARES [68], aiming at the development of a new purpose-built satellite communications system, including low-cost airborne terminals and a new open satellite communication protocol for Air Traffic Management (ATM). The new proposed communication protocol is based on E-SSA in the the re- turn link (aircraft to hub) and is implemented in 200 kHz channels in L-band (1-2 GHz) with a chip rate of 160 kchip/s and a spreading factor SF = 16 or 4 [69, 70]. The performance of the new proposed telecommunication protocol have been verified means of an end-to-end Verification Test Bed (see Fig. 3.3). Laboratory measurements with the verification testbed confirm that a through- put above 1 bit/chip can be achieved with spreading factor SF = 16 (see Fig. 3.4) in the aeronautical channel with a target PLR = 10−3. Slightly lower spec- tral efficiency, around 0.8 bits/s/Hz, is achieved with SF = 4 (see Fig. 3.5). 35
- 50. 36 CHAPTER 3. OUTCOMES (a) Throughput (b) PLR Figure 3.1: Laboratory vs simulation results for CRDSA for Nrep = 4, Nslots=66, QPSK modulation, 3GPP FEC ρ = 1/2, symbol rate Rs = 128 ksymbols/s, packet block size 488 bits, Es/N0 = 10 dB in the presence of lognormal packets power unbalance with mean µ = 0 dB, standard deviation σ = 2 dB.
- 51. 3.1. PROTOTYPES 37 0 0.5 1 1.5 2 2.5 10 -7 10 -6 10 -5 10 -4 10 -3 10 -2 10 -1 10 0 MAC load (bit/s/Hz) Packet Loss Ratio (PLR) !=0dB, analitycal !=2dB, analytical !=3dB, analytical !=0dB, simulation !=2dB, simulation !=3dB, simulation !=0dB, laboratory !=2dB, laboratory !=3dB, laboratory Figure 3.2: PLR vs. MAC loads and diffrent values of power imbalance standard deviation (C/N=-16 dB) [65].
- 52. 38 CHAPTER 3. OUTCOMES It is also important to highlight the very good match between simulation and measurement results. An ACRDA prototype is currently under development by DLR (Germany) under the frame of an ESA contract [71]. Measurement results will be available in 2016. In chapter 2 from [72], Riccardo de Gaundezi and Gennaro Gallinaro com- pare the simulated CRDSA and E-SSA aggregated throughput versus the uncon- strained and constrained in the used code rate and modulation. The simulated CRDSA throughput with 3 replicas, a PLR=10−3, QPSK modulation and 3GPP turbo code with 100 bits block size is compared to the unconstrained and QPSK with FEC 100 bit block size constrained capacity bounds. In case of CRDSA the packets have been using a uniform in dB random packet power distribution with a maximum value of Eb/N0 of 6, 9, 12 and 20 dB respectively. It can be observed that the simulated CRDSA throughput is quite close to the modula- tion and FEC constrained capacity at medium values of Eb/N0. The distance is increasing for large values of Eb/N0 possibly due to the fixed modulation format (QPSK) and code rate (1/3) adopted for CRDSA. For E-SSA, with an average SNR after de-spreading of 10 dB and using a uniform in dB random packet power distribution with a range of 6 dB, an unconstrained capacity bound of 2.5 bits/s/Hz and a constrained capacity bound with BPSK, spreading factor SF = 16 and FEC coding rate ρ = 1/3 of 2 bits/s/Hz are derived. These results show that the simulated E-SSA aggregated throughput of 2 bit/s/Hz presented in Sect. 2.2.1 match the modulation and FEC constrained capacity. 3.2 Standards and Systems In the last years, a number of satellite communication standards have been developed integrating the most recent RA techniques. In the broadband access domain, the second generation of the DVB-RCS standard for interactive satellite systems [73], has added CRDSA as an optional feature, which is particularly advantageous for the SCADA and consumer profiles. In the mobile context, the inclusion of the low-latency profile in the DVB-SH standard for satellite services to handheld [74], enabled the possibility to add a return link (earth station to hub), which is enhancing the DVB-SH standard for the support of interactive services. The ETSI S-band Mobile Interactive Multimedia (S-MIM) standard [75, 65] has adopted the E-SSA technique for the asynchronous access. The E-SSA RA technology is finding commercial application in the recently deployed Eutelsat Broadcast Interactive System (EBIS) (see Fig. 3.6). The core of the EBIS system is the Fixed Satellite Interactive Multimedia (F-SIM) protocol [76]. The F-SIM protocol is implemented by the Smart LNB [77] on the earth stations. The Smart LNB implements conventional DVB-S2 on the forward link (hub to earth station) [78] and a proprietary evolution of the S- MIM standard [75] in the return link (earth station to hub). The EBIS system is already deployed in Ka-band on Eutelsat Ka-Sat satellite, as well as in Ku- band on different Eutelsat satellites.
- 53. 3.2. STANDARDS AND SYSTEMS 39 Figure 3.3: ANTARES communication standard verification testbed.
- 54. 40 CHAPTER 3. OUTCOMES Figure 3.4: SIC performance for ANTARES communication protocol with chip rate Rc = 160 kchip/s, SF = 16, BPSK, code rate ρ = 1/3, data word 512 bits and 3 IC iterations.
- 55. 3.2. STANDARDS AND SYSTEMS 41 Figure 3.5: SIC performance for ANTARES communication protocol with chip rate Rc = 160 kchip/s, SF = 4, BPSK, code rate ρ = 1/3, data word 2048 bits and 3 IC iterations.
- 56. 42 CHAPTER 3. OUTCOMES Figure 3.6: Eutelsat Broadcast Interactive System [76]. 3.3 Example of a M2M satellite network In this section, a realistic example of an M2M satellite network is described with the aim of describing the integration of the new proposed RA schemes with the upper layer protocols and to perform a preliminary system dimensioning exercise based on the traffic characteristics. The selected scenario, is an aeronautical satellite network to track and mon- itor airplanes flying over oceanic airspace (e.g. over the Atlantic). The system uses the L-band (1-2 GHz) for radio links between the airplanes and the satel- lite, with an uplink frequency of 1.55 GHz (aircraft to satellite) and a downlink frequency of 1.65 GHz (satellite to aircraft). Frequency channels on uplink and downlink have a bandwidth of 200 kHz. Aircrafts are equipped with a low gain antenna and low power satellite communications terminal. A hub located on ground communicates with the satellite using a frequency band allocated to Fixed Satellite Services (e.g. in C, Ku or Ka-band). The satellite relays the sig- nals received from the hub to the airplanes (forward link) as well as the signals received from the airplanes to the hub (return link). The hub is connected via a ground network to the to the Air Traffic Control (ATC) centre. A bidirectional communication link is established between the ATC centre and the airplanes. The network architecture is depicted in Fig. 3.7. Airplanes transmit position reports periodically, and also triggered by an event on-board or upon request from the ATC centre. On average, a message is generated every 40 seconds per aircraft with an average length of 250 bytes, resulting in an average bit rate per aircraft of 50 bits/s. Headers from network
- 57. 3.3. EXAMPLE OF A M2M SATELLITE NETWORK 43 Land Ocean Satellite HUB Air Traffic Control L-band C-, Ku- or Ka-band Figure 3.7: Aeronautical Satellite Network for Air Traffic Management. protocols are compressed down to 4-7 bytes, e.g. using ROHC [79, 80], which allows an efficient use of bandwidth on low- and medium-speed links. At the link layer, the Generic Stream Encapsulation (GSE) protocol defined in [81] and adapted to the aeronautical environment in [69]. This encapsulation protocol is characterized by its flexibility and low overhead in the range of 3-4%. If we account for the compressed network headers and encapsulation overhead, the resulting average information bit rate per aircraft at physical layer is 54 bits/s. We have seen in Sect. 3.1 that the ANTARES project proposes an air in- terface that can achieve spectral efficiencies in the return link above 1 bit/chip for a target PLR = 10−3 by using the E-SSA technique. This means that in a 160 kchip/s channel (200 kHz bandwidth) we can track and monitor ≈ 3000 airplanes flying simultaneously over the oceanic airspace. Today, there are less of 3000 aircrafts flying over the Atlantic ocean over a whole day. At a bit rate of 3.3kbit/s per aircraft, the average transmission delay for each message will be 650ms, and the message will be received by the hub in less than 1s (if we account for the propagation delay). In total, the end-to-end message transmission delay from aircraft to Air Traffic Control centre will be < 1.5s, if we account for hub message decoding time and ground network delay from hub to satellite control centre. These are remarkable performance results with a highly aggregated through- put, low PLR and low end-to-end delay, if we consider that a very large network population is sharing a communication channel with a very large propagation delay.
- 61. Chapter 4 Contention Resolution Diversity Slotted Aloha (CRDSA): an Enhanced Random Access Scheme for Satellite Access Packet Networks c 2007 IEEE. Reprinted, with permission, from Enrico Casini, Riccardo De Gau- denzi and Oscar del Rı́o Herrero, Contention Resolution Diversity Slotted Aloha (CRDSA): an Enhanced Random Access Scheme for Satellite Access Packet Net- works, IEEE Transactions on Wireless Communications, Vol. 6, Issue 4, pp. 1408-1419, April 2007 55 citations by end June 2015 (source: Web of Science) 47
- 63. Chapter 5 Generalized Analytical Framework for the Performance Assessment of Slotted Random Access Protocols c 2014 IEEE. Reprinted, with permission, from O. del Rı́o Herrero and R. De Gaudenzi, Generalized Analytical Framework for the Performance Assessment of Slotted Random Access Protocols, IEEE Transactions on Wireless Communications, Vol. 13, Issue 2, pp. 809-821, February 2014 75
- 65. Chapter 6 CRDSA patent: Method of packet mode digital communication over a transmission channel shared by a plurality of users EP1686746 (B1), Transmission de donnees par paquets a travers un canal de transmis- sion partage par plusierus utilisateurs Inventors: E. Casini, O. del Rio Herrero, R. De Gaudenzi, D. M. E. Delaruelle, J.P. Choffray Patent filed: 30 January 2006 Patent application publication: 2 August 2006 Patent published: 16 April 2008 US8094672 (B2), Method of packet mode digital communication over a transmission channel shared by a plurality of users Inventors: E. Casini, O. del Rio Herrero, R. De Gaudenzi, D. M. E. Delaruelle, J.P. Choffray Patent filed: 30 January 2006 Patent application publication: 3 August 2006 Patent published: 10 January 2012 103
- 67. 105
- 68. 106 CHAPTER 6. CRDSA PATENT
- 69. Chapter 7 High Efficiency Satellite Multiple Access Scheme for Machine-to-Machine Communications c 2012 IEEE. Reprinted, with permission, from O. del Rı́o Herrero and R. De Gaudenzi, High Efficiency Satellite Multiple Access Scheme for Machine-to-Machine Communica- tions, IEEE Transactions on Aerospace and Electronic Systems, Vol. 48, Issue 4, pp. 2961-2989, October 2012 107
- 71. Chapter 8 E-SSA patent: Methods, apparatuses and system for asynchronous spread-spectrum communication EP2159926 (A1), Methods, apparatuses and system for asynchronous spread-spectrum communication Inventors: O. del Rio Herrero, R. De Gaudenzi Patent filed: 26 August 2008 Patent application publication: 3 March 2010 US7990874 (B2), Methods, apparatuses and system for asynchronous spread-spectrum communication Inventors: O. del Rio Herrero, R. De Gaudenzi Patent filed: 25 August 2009 Patent application publication: 4 March 2010 Patent published: 2 August 2011 153
- 73. 155
- 74. 156 CHAPTER 8. E-SSA PATENT
- 75. Chapter 9 Asynchronous Contention Resolution Diversity ALOHA: Making CRDSA Truly Asynchronous c 2014 IEEE. Reprinted, with permission, from R. De Gaudenzi, O. del Rı́o Herrero, G. Acar and E. Garrido Barrabés, Asynchronous Contention Resolution Diversity Aloha: Making CRDSA Truly Asynchronous, IEEE Transactions on Wireless Communications, Vol. 13, Issue 11, pp. 6193-6206, November 2014. 157
- 77. Chapter 10 ACRDA patent: Method and apparatus for transmitting data packets over a transmission channel shared by a plurality of users WO2015000518 (A1), Method and apparatus for transmitting data packets over a trans- mission channel shared by a plurality of users Inventors: R. De Gaudenzi, O. del Rio Herrero Patent filed: 3 July 2013 Patent application publication: 8 January 2015 187
- 79. 189
- 83. Chapter 11 Conclusions This thesis has presented three novel random access techniques that offer an excellent performance over the satellite environment and are well suited to the type of applica- tions and services that have been recently emerging (e.g. consumer broadband access, interactive TV, machine-to-machine communications, transaction and safety of life com- munications). The first RA technique, CRDSA, is suited for slotted access (e.g. TDMA or MF-TDMA) that is typically used in broadband access networks. The second and third techniques, E-SSA and ACRDA, are fully asynchronous and are better suited for low cost fixed and mobile satellite messaging systems. The main difference between E-SSA and ACRDA is that E-SSA is a spread spectrum based RA technique. The three techniques have been analyzed in depth in this thesis, covering a description of the tech- niques, their key features, the processing at the transmitter and receiver sides, together with implementation aspects. The new RA techniques can achieve an aggregated throughput above 1 bit/s/Hz for a PLR < 10−3 , which is typically required in data communications for an efficient per- formance of the upper layers network protocols. This represents a major performance improvement compared to known RA techniques currently used by satellite communi- cation systems such as S-ALOHA (factor 1000), DSA (factor 50) and SSA (factor 2). The performance of the new schemes is further boosted in the presence of packets power imbalance at the receiver side, which is typically present in the satellite communica- tions channel if no closed loop power control mechanisms are implemented. CRDSA and ACRDA can achieve a throughout in the order of 1.4 and 1.5 bit/s/Hz respectively in the presence of power imbalance. Particularly remarkable is the performance of E- SSA in front of power imbalance, which can achieve a throughput of 2 bit/s/Hz for a PLR < 10−3 . Power imbalance is not only beneficial for the performance of the novel RA schemes presented in this thesis, but also reduces complexity eliminating the need for closed loop power control schemes. On the contrary, the performance improvement of S-ALOHA and DSA in the presence of power imbalance is rather limited (power cap- ture effect), and highly destructive for SSA (i.e. throughput reduction from 0.5 down to 0.02 bits/s/Hz for a target PLR < 10−3 ). By comparing the performance improvement of E-SSA vs. SSA in front of power imbalance we get a factor 100. In terms of delay, E-SSA and ACRDA achieve similar delays to conventional RA schemes, while CRDSA introduces a higher delay due to the framed nature of the scheme (i.e. all packet replicas have to be transmitted within a common frame, and therefore, the earth station has to wait until the start of the next RA frame to transmit all packet replicas). As result, reducing the frame duration is desirable in CRDSA. The presented RA schemes use iterative interference cancelation at the hub demod- ulator to resolve the collisions. To achieve high throughput with high reliability in RA, 193
- 84. 194 CHAPTER 11. CONCLUSIONS several design aspects at system, physical and MAC layer have been considered alto- gether. The performance of RA schemes exploiting IC are particularly sensitive to the choice of a number of aspects such as the number of replicas for each packet (CRDSA, ACRDA), number of CDMA codes (E-SSA), the channel coding rate, the FEC code performance at low SNR and, in the presence of colliding packets, the packets power imbalance received at the hub and the packet detection and channel estimation perfor- mances. This thesis, has performed detailed analyses and simulations for the most important design parameters. CRDSA best performance is achieved when 3 replicas are used and a channel coding rate ρ = 1/3. The use of 3 replicas, as opposed to 2, helps eliminating the ”loop” problem present in CRDSA and the coding rate ρ = 1/3 as opposed to 1/2 allows to successfully decode a packet even in the presence of one collision which helps in triggering the iterative IC process at the demodulator. A random access frame with more than 64 slots and 5 IC iterations are typically required to achieve good performance. On the physical layer side, one important aspect in CRDSA is channel estimation performance. This is done in a two step approach, with a first detection on the burst preamble for packet decoding, and once the packet bits have been successfully decoded, a second data aided channel estimation using also the payload symbols. It has been shown that this second step data aided channel estimation achieves IC results very close to ideal channel estimation. For E-SSA, important design parameters are the sliding window size, the channel coding rate, the spreading factor, the number of CDMA codes, the number of IC itera- tions and the packet detection and channel estimation performances. It has been shown that a window size W = 3 packets gives good performance, as it covers the span of the interfering packets over the packet being decoded. Low coding rates are important in RA as we need to decode packets at very low SNIR in order to trigger the iterative IC process. It has been shown that FEC codes with rate ρ = 1/3 and small packets sizes (e.g. 100 information bits) have reasonable FER performance with Eb/N0 = −0.5 dB. It has been shown that the spreading factor drives the level of traffic aggregation, which is higher when large spreading factors are used (e.g. SF = 256). It has been shown that with large spreading factors the interference level is almost constant over the whole packet duration and can be approximated as AWGN. The code collision probabilities has been analyzed in a fully asynchronous scenario and shown that 1 code is sufficient to achieve code collision probabilities well below the target PLR < 10−3 . Regarding the number of IC iterations, it has been shown that only 3 IC iterations are required at each sliding window position to get the best performance. Packet detection is a key aspect in E-SSA, as we need to operate in front of very high interference levels (e.g. Ec/(N0 + I0) ' −30 dB). Results obtained show that a preamble length of 128 symbols can achieve a probability of miss detection and false alarm below 10−3 . Channel estimation is performed in two steps as for CRDSA yielding a very good IC efficiency close to ideal channel estimation. The optimal design parameters for ACRDA are similar to those of CRDSA and E-SSA, as this technique combines elements from them. It is important to remark, that in ACRDA the optimal number of replicas is 2 instead of 3 as found for CRDSA. Due to the asynchronous nature of the ACRDA scheme, the probability of loops is highly reduced already with two replicas, and there is no need to increase further complexity at the demodulator with one additional replica. The optimal sliding window size W is 3 virtual frames following the same rationale as for E-SSA. This thesis has also presented novel analytical models to study the throughput and PLR performance of the new RA schemes. The proposed analytical model allows to predict the performance of the RA schemes in the presence of arbitrary FEC schemes, packet size, traffic and received packets power distributions. The model takes into
- 85. 195 account the fact that thanks to powerful physical layer coding, packets may be decoded with non zero probability also in the presence of collisions. The model also considers the effect of slot time diversity in case more than one packet replica is transmitted in the same frame, as it is the case for DSA. Two types of IC strategies have been modeled: successive interference cancellation over one slot position (CRDSA) or one window position (E-SSA and ACRDA), and iterative interference cancellation across the slots due to successful reception of packet replicas in other slots (CRDSA and ACRDA). The new model have been used to perform sensitivity analyses against a number of design parameters for the new RA schemes presented in this thesis, and to compare the performance improvement against conventional RA schemes. The good match between analytical and simulation results confirm the validity of the models, and make them extensible for future use in new or evolved RA schemes. The new RA schemes have triggered the interest from industry who have developed real time laboratory prototypes [63, 67, 71] and validated the simulation results. In ad- dition, CRDSA and E-SSA have been incorporated into two satellite telecommunication standards confirming the interest from industry [73, 75]. Finally, we conclude this thesis by mentioning that the E-SSA RA technology is finding commercial application in the recently deployed Eutelsat Broadcast Interactive System [76].
- 87. Chapter 12 Future work The development of the new RA techniques presented in this thesis has opened up a new field of research. In this chapter some of the key open areas of research related to high performance satellite RA are listed. The choice of the physical layer Forward Error Correcting (FEC) scheme for CRDSA is an important design driver. In CRDSA we are not operating in AWGN-like condi- tions, but in the presence of heavy co-channel interference the FEC collision resolution capability is important. Under these conditions of heavy multiple access interference, it is important to choose a FEC code that is able to recover a few packets and will then trigger the iterative IC process. Therefore, lower coding rates, although apparently re- ducing the individual packet bit rate, may be important to enhance the aggregated RA scheme performance. The final choice of coding rate is a tradeoff between increased overhead and RA performance improvement and has been preliminary studied in [60]. Regarding the FEC block size, it is preferable to have a small block size as this provides a smaller PLR in the lower signal-to-noise ratio (SNR) range [60]. Again, the reasoning is that successful detection of few packets under heavy co-channel interference conditions will be able to trigger the iterative IC process and eventually recover all packets from the RA frame. As a general rule, low coding rates (≤ 1/2) and small packet sizes (≤ 1000 bits) are recommended for CRDSA and in general, all RA schemes implementing IC. Another interesting subject of research is related to reduction of the transmitter power consumption and cost. The use of constant envelope modulation like Continuous Phase Modulation (CPM) can bring several advantages, as it allows to replace the transmitter frequency up-conversion unit with a cheaper nonlinear frequency multiplier. Furthermore the use of CPM modulation allows to operate the transmitter amplifier in highly nonlinear mode with consequent reduction of the DC power requirements. Past work has shown the possibility to combine CPM with DS-SS [149]. Although CPM has been used in commercial S-ALOHA and SSA systems such as Sat3Play [150] and ArcLight [151], its use for iterative SIC based RA is still under investigation. While for SSA and E-SSA the use of low-order modulation such as BPSK with low coding rate is close to optimum, recent investigations [152] indicate that for non-SS RA there may be some throughput advantage using modulation orders above QPSK when the SNR is above a certain value. Using QPSK in E-SSA does not bring any throughput advantage, but just an increase on the sensitivity to channel estimation phase errors. The use of a pilot in quadrature allows to have a continuous phase estimation which provides robustness in case of channels with phase noise and/or fading (mobile environment). However, concerning CRDSA, the use of QPSK instead of BPSK is related to the achievable spectral efficiency considering the TDMA type of access scheme. The idea is to use an energy efficient signaling scheme like QPSK (or a higher order 197
- 88. 198 CHAPTER 12. FUTURE WORK modulation) with low rate FEC to minimize the power of the transmitter while getting very good aggregate RA throughput. When the RA network allows to exploit the incoming packet power imbalance, this power distribution can be optimized to enhance the system throughput. While for E- SSA RA some semi-analytical optimization has been derived in [153], a similar effort for non spread spectrum RA is still undergoing. Initial findings indicate that also for CRDSA and ACRDA a uniform in dB incoming packet distribution can improve the throughput performance. In systems with a high SNR, but reduced packet power dynamic range, MMSE preprocessing may be implemented to enhance the SNIR of the received signal before the signal demodulation. The MMSE detector can exploit the knowledge of the signal structure to compute the matrix filter. Preliminary analyses and simulation results on a newly proposed ME-SSA scheme can be found in [154]. A major drawback of TDMA is its high peak transmit power requirement resulting from the fact that each terminal is transmitting on a small fraction (1/Nslots) of the frame duration. Since the PLR of a system is directly related to the ratio between the bit energy and the noise floor level (i.e. the Eb/N0), TDMA needs a (peak) power transmission Nslots times higher than FDMA or CDMA. In thin route satellite commu- nications networks Multi-Frequency TDMA (MF-TDMA) was introduced to mitigate this problem. The multi-frequency concept can be easily applied to CRDSA or ACRDA leading to MF-CRDSA [58]. The peak power gain reduction as for MF-CRDSA will be proportional to the ratio of frequency to time slots of the system. Besides, using the frequency domain will reduce the RA frame duration thus reducing the end-to-end delay in CRDSA. The multi-frequency concept can be also applied in ACRDA to reduce the terminal peak power requirements. The exploitation in ACRDA of techniques such as soft combining of packet replicas proposed in [140] or the combining of the best replica packets chunks to generate a better packet of interest as suggested in [143] or the Coded Slotted ALOHA soft combining scheme [136] represent potential area of future investigation. Similarly one can consider the techniques proposed in [139], [140] to pseudo-randomize the slot positions as a function of the transmitted data to remove the replica signaling overhead. This is particularly interesting in case small packets have to be transmitted. The issue of advanced Spread Spectrum RA demodulation in case of multi-beam satellite networks has been recently tackled in [155]. It has been shown that for a mobile multi-beam system exploiting the S-MIM standard further RA capacity gain can be achieved using full frequency reuse among the beams and joint beam processing at the gateway. In particular, two different multi-beam joint processing techniques have been investigated by simulation: basic (decoded packets are send to upper layers belonging to other beams) and enhanced cooperation (the decoded data is sent to the co-frequency demodulators on top the basic cooperation). The latter approach, although more complex to implement, brings extra gain on top of the basic cooperation. More work in this area is required to assess the practical feasibility and gains achievable exploiting joint beam processing. In this thesis we have shown that RA alone can achieve throughput performance truly comparable with DAMA schemes with advantages in terms of transmission latency and signaling reduction. Furthermore, the presented advanced RA techniques are easily scalable to very large networks. However, there may be applications where the type of traffic can be both bursty or bulky. In this case the combination of high performance RA schemes with DAMA can represent the best solution to cope with this very time variant traffic conditions. More work is required to investigate the best way to adaptively switch from the new efficient and reliable RA to DAMA type of access schemes [144, 156]. Finally, the techniques presented in this thesis can also be adopted in terrestrial
- 89. 199 wireless networks to improve their performance in terms of capacity and energy con- sumption. Examples systems for application are Radio Frequency Identification (RFID) systems [139, 140] and M2M networks [157, 158].
- 91. Bibliography [1] M. Williamson, Spacecraft Technology: The Early Years, The Institution of Engi- neering and Technology, 2006. [2] K.M.S. Murthy, J. Alan, J. Barry, B.G. Evans, N. Miller, R. Mullinax, P. Noble, B. O’Neal, J.J. Sanchez, N. Seshagiri, D. Shanley, J. Stratigos, J.W. Warner, VSAT user network examples, IEEE Communications Magazine, Vol. 27, Issue 5, pp. 50- 57, May 1989. [3] B. Arbesser-Rastburg, R. Bellini, F. Coromina, R. De Gaudenzi, O. del Rio Her- rero, M. Hollreiser, R. Rinaldo, P. Rinous, A. Roederer, R&D directions for next generation broadband multimedia systems: an ESA perspective, 20th AIAA Interna- tional Communications Satellite Systems Conference (ICSSC), Montreal, Canada, May 13-15, 2002. [4] Satellite Industry Association State of Satellite Industry Report, September 2014. [5] Euroconsult, High throughput satellites: on course for new horizons, 2014 Edition. [6] Northern Sky Research, HTS Revenues to Overtake FSS in Battleground Applica- tions, http://www.nsr.com/news-resources/nsr-in-the-press/nsr-press-releases/hts- revenues-to-overtake-fss-in-battleground-applications/. [7] Via Satellite, The Continuing Momentum of M2M/IoT via Satellite, http://www.satellitetoday.com/publications/2015/01/20/the-continuing- momentum-of-m2miot-via-satellite/. [8] G. Maral, M. Bousquet, Satellite Communicaitons Systems, Fifth Edition, John Wiley & Sons Ltd, 2009. [9] ETSI EN 300 421 Ver. 1.1.2, Digital Video Broadcasting (DVB); Framing structure, channel coding and modulation for 11/12 GHz satellite services, August 1997. [10] ETSI EN 302 307-1 Ver. 1.4.1, Digital Video Broadcasting (DVB); Second gener- ation framing structure, channel coding and modulation systems for Broadcasting, Interactive Services, News Gathering and other broadband satellite applications; Part 1: DVB-S2, November 2014. [11] G. Maral, VSAT Networks, Second Edition, John Wiley & Sons Ltd, 2003. [12] A.H. Rana, J.S. McCoskey, W.A. Check, VSAT technology, trends, and applica- tions, Proceedings of the IEEE, Vol. 78, Issue 7, pp. 1087 - 1095, July 1990. [13] D. Chakraborty, VSAT Communications Networks - An overview, IEEE Commu- nications Magazine, Vol. 26, No. 5, pp. 10-24, May 1988. [14] F.A. Tobagi, Multiaccess Protocols in Packet Communication Systems, IEEE Trans- actions on Communications, Vol. COM-28, No. 4, pp. 468-488, April 1980. 201
- 92. 202 BIBLIOGRAPHY [15] C. Retnadhas, Satellite Multiple Access Protocols, IEEE Communications Maga- zine, Vol. 18, No. 5, pp. 16-20, September 1980. [16] C.J. Wolejsza, D. Taylor, M. Grossman, W.P. Osborne, Multiple Access Protocols for Data Communications via VSAT Networks, IEEE Communications Magazine, Vol. 25, No. 7, July 1987. [17] D. Raychaudhuri, K. Joseph, Channel Access Protocols for Ku-band VSAT Net- works: a comparative evaluation, IEEE Communications Magazine, Vol. 26, No. 5, pp. 34-44, May 1988. [18] H. Peyravi, Medium Access Control Protocol Performance in Satellite Communica- tions, IEEE Communications Magazine, Vol. 37, No. 3, pp. 62-71, March 1999. [19] P.M. Feldman, An Overview and Comparison of Demand Assignment Multiple Ac- cess (DAMA) Concepts for Satellite Communications Networks, RAND corpora- tion, 1996. [20] J.G. Puente, A PCM-FDMA demand assignment satellite multiple access experi- ment, AIAA 2nd Communications Satellite Systems Conference, paper No. 68-451, 8-10 April 1968, San Francisco (USA). [21] G. Frenkel, The grade of service in multiple-access satellite communications systems with demand assignments, IEEE Transactions on Communications, Vol. 22, No. 10, pp. 1681-1685, October 1974. [22] W.E. Leland, On the self-similar nature of Ethernet traffic (Extended version), IEEE/ACM Transactions on Networking, Vol. 2, No. 1, February 1994. [23] V. Paxson, S. Floyd, Wide Area Traffic: the failure of Poisson modeling, IEEE/ACM Transactions on Networking, Vol. 3, No. 3, June 1995. [24] I. Jacobs, L.N. Lee, A. Viterbi, R. Binder, R. Bressler, N.T. Hsu, R. Weissler, CPODA - A demand Assignment protocol for Satnet, in Proceeding of the fifth symposium on data communications (SIGCOMM’77), pp. 2.5-2.9, 1977. [25] T. Le-Ngoc, S.V. Krishnamurthy, Performance of combined free/demand assign- ment multiple-access schemes in satellite communications, Wiley International jour- nal of satellite communications, Vol. 14, No. 1, pp. 11-21, January 1996. [26] T. Le-Ngoc and I.M. Jahangir, Performance analysis of CFDAMA-PB protocol for packet satellite communications, IEEE Transactions on Communications, Vol. 46, No. 9, pp. 1206-1214, September 1998. [27] O. del Rı́o Herrero, G. Foti and G. Gallinaro, Spread-spectrum techniques for the provision of packet access on the reverse link of next-generation broadband multi- media satellite systems, IEEE Journal on Selected Areas in Communications, Vol. 22, Issue 3, pp. 574 - 583, April 2004 [28] Z. Jiang, Y. Li, V.C.M. Leung, A Predictive Demand Assignment Multiple Access Protocol for Broadband Satellite Networks Supporting Internet Applications, IEEE International Conference on Communications (ICC), pp. 2973 - 2977, Vol.5, 28 April - 2 May 2002, New York (USA) [29] M. Hasegawa, G. Wu, M. Mizuno, Applications of nonlinear prediction methods to the Internet traffic, Proc. International Symposium on Circuits and Systems, vol. 3, pp. 169-172, May 2001, Sydney (Australia).
- 93. BIBLIOGRAPHY 203 [30] H.W. Lee and J. W.Mark, Combined random/reservation access for packet switched transmission over a satellite with on-board processing: Part I Global beam satellite, IEEE Transactions on Communications, Vol. COM-31, No. 10, pp. 1161-1171, Oc- tober 1983. [31] D. M. Chitre, VSAT Networks: Architectures, Protocols, and Management, IEEE Communications Magazine, July 1988, Vol. 26, No. 7, pp. 28-38, July 1988. [32] D. Bertsekas and R. Gallager, Data Networks, Second Edition, Prentice Hall, 1992. [33] S. Keshav, An Engineering Approach to Computer Networking, Addison-Wesley Professional, 1997. [34] L. Kleinrock and F.A. Tobagi, Packet switching in radio channels: part I - Car- rier sense multiple access modes and their throughput-delay characteristics, IEEE Transactions on Communications, Vol. COM-23, No. 12, pp. 1400-1416, December 1975. [35] R. M. Metcalfe and D.R. Boggs, ETHERNET: Distributed packet switching for local computer networks, Communications of the ACM magazine, Vol. 19, No. 7, pp. 395-404, July 1976. [36] F.A. Tobagi and L. Kleinrock, Packet Switching in Radio Channels: PartII - The Hidden Terminal Problem in Carrier Sense Multiple-Access and the Busy-Tone Solution, IEEE Transactions on Communications, Vol. COM-23, No. 12, December 1975. [37] IEEE Std 802.3-2005, IEEE Standard for Information technology; Telecommunica- tions and information exchange between systems local and metropolitan area net- works; Specific requirements Part 3: Carrier sense multiple access with collision detection (CSMA/CD) access method and physical layer specifications, December 2005. [38] IEEE Std 802.11-2012, IEEE Standard for Information technology; Telecommu- nications and information exchange between systems local and metropolitan area networks; Specific requirements Part 11: Wireless LAN Medium Access Control (MAC) and Physical Layer (PHY) Specifications, March 2012. [39] P. Karn, MACA - A New Channel Access Method for Packet Radio, Proceedings of the ARRL/CRRL Amateur Radio 9th Computer Networking Conference, pp. 134-140, 22 September 1990, London and Ontario (Canada). [40] V. Bharghavan, A. Demers, S. Shenker, L. Zhang, MACAW: A Media Access Pro- tocol for Wireless LAN’s, ACM SIGCOMM, 1994. [41] C. Fullmer, J.J. Garcia-Luna-Aceves, Floor Acquisition Multiple Access (FAMA) for packet radio networks, Computer Communication Review, Vol. 25, No. 4, ACM SIGCOMM ’95, Cambridge, MA, USA, 28 Aug.-1 Sept. 1995. [42] 3GPP TS 25.214 v3.12.0, Physical Layer Procedures (FDD); Release 1999, March 2003. [43] N. Abramson, The ALOHA system - Another alternative for computer communi- cations, in 1970 Fall Joint Comput. Conf. AFIPS Conf. Proc., vol. 37, pp. 281-285, AFIPS Press, Montvale (NJ), 1970. [44] N. Abramson, The Throughput of Packet Broadcasting Channels, IEEE Transac- tions on Communications, Vol. COM-25, No. 1, pp. 117-128, January 1977.
- 94. 204 BIBLIOGRAPHY [45] L. G. Roberts, ALOHA Packet Systems with and Without Slots and Capture, Arpanet Satellite System Note 8 (NIC Document #11290), 26 Jun. 1972. [46] G.L. Choudhury and S.S. Rappaport, Diversity ALOHA - A Random Access Scheme for Satellite Communications, IEEE Transactions on Communications, Vol. COM-31, pp. 450-457, March 1983. [47] O. del Rı́o Herrero and R. De Gaudenzi, Generalized Analytical Framework for the Performance Assessment of Slotted Random Access Protocols, IEEE Transactions on Wireless Communications, Vol. 13, No. 2, pp. 809-821, February 2014. [48] N. Abramson, Multiple Access in Wireless Digital Networks, IEEE Proceedings, Vol. 82, No. 9, pp. 1360-1370, September 1994. [49] C. Pateros, Novel direct sequence spread spectrum multiple access technique, in MILCOM 2000, 21st Century Military Communications Conference Proceedings, Vol. 2, pp. 564-568, Los Angeles (CA) Oct. 22-25, 2000. [50] O. del Rı́o Herrero and R. De Gaudenzi, High Efficiency Satellite Multiple Ac- cess Scheme for Machine-to-Machine Communications, IEEE Transactions on Aerospace and Electronic Systems, Vol. 48, No. 4, pp. 2961-2989, October 2012. [51] ETSI EN 301 790 v1.5.1, Digital Video Broadcasting (DVB); Interaction channel for satellite distribution systems, May 2009. [52] Telecommunication Industry Association TIA-1008-A, IP Over Satellite, May. 2006. [53] S. Verdu, Multiuser detection, Cambridge Univ. Press, 1998. [54] C. C. Lee, Random signal levels for channel access in packet broadcast networks, IEEE J. Select. Areas Commun., Vol. 5, No. 6, pp. 1026-1034, July 1987. [55] I. Cidon, J. Kodesh and M. Sidi, em Erasure, capture, and random power level selection in multiple access systems, IEEE Transactions on Communications, Vol. 36, No. 3, pp. 263-271, March 1988. [56] R. O. LaMaire, A. Krishna and M. Zorzi, On the randomization of transmitter power levels to increase throughput in multiple access radio systems, Journal Wireless Networks, Vol. 4, No. 3, pp. 263 - 277, March 1998, Kluwer Academic Publishers Hingham, MA (USA). [57] E. Casini, R. De Gaudenzi, O. del Rı́o Herrero, Contention Resolution Diversity Slotted Aloha (CRDSA): an Enhanced Random Access Scheme for Satellite Access Packet Networks, IEEE Transactions on Wireless Communications, Vol. 6, No. 4, pp. 1408-1419, Apr. 2007. [58] E. Casini, O. del Rı́o Herrero, R. De Gaudenzi, D.M.E. Delaruelle, J.P. Choffray, Transmission de donnees par paquets a travers un canal de transmission partage par plusierus utilisateurs, European Patent EP1686746 (B1), 30 January 2006. [59] O. del Rı́o Herrero, R. De Gaudenzi, Methods, apparatuses and system for asyn- chronous spread-spectrum communications, U.S. Patent 7,990,874 B2, August 2, 2011. [60] O. del Rı́o Herrero, R. De Gaudenzi, J. Pijoan Vidal, Design Guidelines for Ad- vanced Random Access Protocols, 30st AIAA International Communications Satel- lite Systems Conference (ICSSC), 24-27 Sep. 2012, Ottawa (Canada). [61] R. De Gaudenzi, O. del Rı́o Herrero, G. Acar, E. Garrido Barrabés, Asynchronous Contention Resolution Diversity Aloha: Making CRDSA Truly Asynchronous, IEEE Transactions on Wireless Communications, Vol. 13, No. 11, pp. 6193-6206, November 2014.
- 95. BIBLIOGRAPHY 205 [62] R. De Gaudenzi, O. del Rı́o Herrero, Method and apparatus for transmitting data packets over a transmission channel shared by a plurality of users, World Intellectual Property Organization Patent WO2015000518 (A1), 3 July 2013. [63] Space Engineering, Advanced Modem Prototype for Interactive Satellite Terminals (AMPIST), ESA Contract No. 20668/07/NL/JD, https://artes.esa.int/advanced- modem-prototype-interactive-satellite-terminals-ampist [64] R.Romanato, E. Rossini, G. Chiassarini, A. Masci, D. Silvi, P. Burzigotti, D. Gi- ancristofaro, AMPIST: An Advanced Modem Prototype For Interactive Satellite Terminals with Contention Resolution Diversity Slotted ALOHA (CRDSA) Capa- bilities, in 17th Ka and Broadband Communications, Navigation and Earth Obser- vation Conference, 3-5 October 2011, Palermo (Italy) [65] S. Scalise, C.P. Niebla, R. De Gaudenzi, O. del Rio Herrero, D. Finocchiaro, A. Arcidiacono, S-MIM: a novel radio interface for efficient messaging services over satellite, IEEE Communications Magazine, Vol. 51 , No. 3, pp. 119-125, March 2013. [66] F. Basile, G. Mendola, Satellite Hub Communication System GPU Based, GPU Technology Conference, 14-17 May 2012, San Jose (CA). [67] M. Andreacci and G. Mendola, F. Collard, D. Finocchiaro, A. Recchia, Enhanced spread spectrum ALOHA demodulator implementation, laboratory tests and satellite validation, Wiley International Journal on Satellite Communications and Network- ing, Special Issue on S-band Mobile Interactive Multimedia, Vol. 32, No. 6, Page 521-533, November/December 2014. [68] AeroNauTicAl REsources Satellite based (ANTARES), ESA Contract 22851/09/NL/AD, https://artes.esa.int/projects/antares [69] ANTARES Communication Standard Technical Specifications, https://artes.esa.int/news/iris-antares-study-updates-draft-satellite- communications-standard [70] J. Batlle, I. Vega, J. Vila, D. Fernandez, L. Albiol, J.M. Cebrian, C. Ciancarelli, P. Burzigotti, New satellite-based communication standard for Air Traffic Manage- ment, in 19th Ka and Broadband Communications, Navigation and Earth Obser- vation Conference and 31st AIAA International Communications Satellite Systems Conference (ICSSC), 14-17 October 2013, Florence (Italy). [71] Prototypes for Advanced Access Schemes, ESA Contract 4000110938/14/NL/CLP, 2014. [72] O. del Rı́o Herrero, R. De Gaudenzi, G. Gallinaro, High-performance random ac- cess schemes, in Co-operative and cognitive satellite systems, S. Chatzinotas, B. Ottersten, R. De Gaudenzi, Ed. Academic Press, 2015, pp. 35-82. [73] ETSI EN 301 542-2 v1.1.1, Digital Video Broadcasting (DVB); Second Generation DVB Interactive Satellite System (DVB-RCS2); Part 2: Lower Layers for Satellite Standard, January 2012. [74] ETSI EN 302 583 v1.1.1, Digital Video Broadcasting (DVB); Framing Structure, channel coding and modulation for Satellite Services to Handheld devices (SH) below 3 GHz, March 2008. [75] ETSI TS 102 721-3 v1.1.1, Satellite Earth Stations and Systems; Air Interface for S- band Mobile Interactive Multimedia (S-MIM); Part 3: Physical Layer Specification, Return Link Asynchronous Access, December 2011.
- 96. 206 BIBLIOGRAPHY [76] A. Arcidiacono, D. Finocchiaro, F. Collard, S. Scalise, F. Lazaro Blasco, R. De Gaudenzi, S. Cioni, N. Alagha, M. Andrenacci, From S-MIM to F-SIM: Making Satellite Interactivity Affordable at Ku and Ka-band, Accepted for Publication in Wiley International Journal on Satellite Communications and Networking, special issue on Ka-band satellite communications, 2015. [77] Eutelsat, Eutelsats Smart LNB for Direct-to-Home antenna, http://www.eutelsat.com/en/services/broadcast/direct-to-home/smart-LNB.html [78] ETSI EN 302 207 v1.3.1, Digital Video Broadcasting (DVB); Second generation framing structure, channel coding and modulation systems for Broadcasting, Inter- active Services, News Gathering and other broadband satellite applications (DVB- S2), March 2013. [79] IETF RFC 5795, The RObust Header Compression (ROHC) Framework, March 2010. [80] IETF RFC 3095, RObust Header Compression (ROHC): Framework and four pro- files: RTP, UDP, ESP, and uncompressed, July 2001. [81] J. Cantillo, B. Collini-Nocker, U. De Bie, O. del Rio, G. Fairhurst, A. Jahn R. Rinaldo,GSE: A Flexible, yet Efficient Encapsulation for IP over DVB-S2 Con- tinuous Generic Streams, International Journal of Satellite Communications and Networking (IJSCN), vol. 26, Issue 3, pp. 231-250, May/Jun. 2008. [82] T. Le-Ngoc, and J.I. Mohammed, Combined free/demand assignment multiple ac- cess (CFDAMA) protocols for packet satellite communications; Universal Personal Communications, in Conf. Rec. 2nd International Conference on Personal Com- munications, 12-15 Oct. 1993, vol. 2, pp. 824 - 828. [83] S. S. Lam, A carrier sense multiple access protocol for local networks, Computer Networks, vol. 4, pp. 21-32, 1980. [84] D. Raychaudhuri, Aloha with Multipacket Messages and ARQ-type Retransmission Protocols-Throughput Analysis, IEEE Trans. on Comm., vol. COM-32, no. 2, pp. 148-154, Feb. 1984. [85] D. Raychaudhuri, Stability, Throughput, and Delay of Asynchronous Selective Re- ject Aloha, IEEE Trans. on Comm., Vol. 35, No. 7, pp. 767-772, July 1987. [86] European Telecommunication Standardisation Institute (ETSI) TR 101 790 V1.2.1, Digital Video Broadcasting (DVB); Interaction channel for Satellite Distribution Systems; Guidelines for the use of EN 301 790, January 2003. [87] P. Patel, and J. Holtzman, Analysis of a simple successive interference cancellation scheme in a DS/CDMA system, IEEE Journal on Selected Areas in Comm., Vol. 12 , No. 5, pp. 796 - 807, June 1994. [88] R. De Gaudenzi, F. Giannetti, and M. Luise, Signal Recognition and Signature Code Acquisition in CDMA Receivers for Mobile Communications, IEEE Transactions on Vehicular Tech., Vol. 47, No. 1, pp. 196 - 208, Feb. 1998. [89] European Telecommunication Standardisation Institute (ETSI) TR 101 112 v3.2.0, Universal Mobile Telecommunications System (UMTS); Selection procedures for the choice of radio transmission technologies of the UMTS, April 1998. [90] R. De Gaudenzi, O. del Rı́o Herrero, Advances in Random Access protocols for satellite networks, International Workshop on Satellite and Space Communications (IWSSC), pp. 331-336, 9-11 September 2009, Siena (Italy).
- 97. BIBLIOGRAPHY 207 [91] A.H. Abdelmonem, T.N. Saadawi, Performance analysis of spread spectrum packet radio network with channel load sensing, IEEE Journal on Selected Areas in Com- munications, Vol. 7, No. 1, January 1989, pp. 161-166. [92] O. del Rı́o Herrero, R. De Gaudenzi, A High-Performance MAC Protocol for Consumer Broadband Satellite Systems, Proceedings of 27th AIAA International Communications Satellite Systems Conference )ICSSC), 1-4 June 2009, Edinburgh (United Kingdom). [93] K. S. Ghilousen, I. M. Jacobs, R. Padovani, A. J. Vitebi, L. A. Weaver, C. W. Wheatley III, On the Capacity of a Cellular CDMA System, IEEE Transactions on Vehicular Technology, Vol. 40. No. 2, pp. 303-312, May 1991. [94] R. De Gaudenzi, F. Giannetti, Analysis of an Advanced Satellite Digital Audio Broadcasting System and Complementary Terrestrial Gap-Filler Single Frequency Network, IEEE Transactions on Vehicular Technology Vol. 42, No. 2, pp. 194-210, May 1994. [95] P. Cardieri, T. S. Rappaport, Statistics of the Sum of Lognormal Variables in Wire- less Communications, Vehicular Technology Conference 2000, Proceedings of the 51st Vehicular Technology Conference Spring, Volume: 3, pp. 1823-1827, May 2000, Tokyo (Japan). [96] 3GPP TS 25.212 v3.11.0, Multiplexing and channel coding (FDD); Release 1999, September 2002. [97] A. Giorgetti, M. Chiani, Influence of Fading on the Gaussian Approximation for BPSK and QPSK With Asynchronous Cochannel Interference, IEEE Transactions on Wireless Comm., Vol. 4, No. 2, pp. 384-388, March 2005. [98] J.G. Proakis, Digital Communications, Fourth Ed. New York, McGraw-Hill, 2001. [99] D.H. Layer, Digital radio takes to the road, IEEE Spectrum, Vol. 38, No. 7, pp. 40-46, July 2001. [100] C.Y. Yeun, Mobile TV technologies, in the Proc. of the 5th International Con- ference on Information and Communications Technology (ICICT), pp. 2-9, 16-18 December 2007, Cairo (Egypt). [101] 3GPP TR 22.868 v8.0.0, Study on Facilitating Machine to Machine Communica- tion in 3GPP Systems; Release 8, March 2007. [102] European Telecommunication Standardisation Institute (ETSI), Document Draft ETSI TS 102 689 v0.3.1, Machine-to-Machine communications (M2M); M2M ser- vice requirements, 2009. [103] C. Pateros, Novel direct sequence spread spectrum multiple access technique, MIL- COM 2000, 21st Century Military Communications Conference Proceedings, 22-25 October 2000, Los Angeles (USA) [104] H. Holma, A. Toskala, WCDMA for UMTS, John Wiley & Sons, Ltd, 2005. [105] T. Le-Ngoc, V. Leung, P. Takats and P. Garland, Interactive Multimedia Satellite Access Communications, IEEE Communications Magazine, Vol. 41, No. 7, pp. 78- 85, July 2003. [106] C.E. Wheatley III, R. Padovani, E. Zehavi, Fast forward link power control in a code division multiple access system, United States Patent No. 5461639, 24 October 1995.
- 98. 208 BIBLIOGRAPHY [107] K.S. Gilhousen, R. Padovani, C.E. Wheatley III, Spread spectrum transmitter power control method and system, United States Patent No. 5,257,283, 26 October 1993. [108] A.M. Monk, L.B. Milstein, Open-loop power control error in a land mobile satellite system, IEEE Journal on Selected Areas in Communications, Vol. 13, No. 2, pp. 205-212, February 1995. [109] D. Boudreau, G. Caire, G. E. Corazza, R. De Gaudenzi, G. Gallinaro, M. Luglio, R. Lyons, J. Romero-Garcia, A. Vernucci, H. Widmer, Wideband CDMA for the Satellite Component ofUMTS/IMT-2000, IEEE Transactions on Vehicular Tech- nology, Vol. 51, No. 2, pp 306-330, Mar. 2002. [110] N. Abramson, Spread aloha CDMA data communications, United States Patent No. 5537397, 16 July 1996. [111] Xiang Feng, Yan Li, Guangguo Bi, A CDMA-slotted ALOHA broadband system for multiservices, IEEE 1998 International Conference on Universal Personal Com- munications (ICUPC’98), Vol. 2, pp. 1131-1135, 5-9 October 1998, Florence (Italy). [112] Y. Tadokoro, H. Okada, T. Yamazato, A. Katayama, A. Ogawa, A new packet detection scheme in CDMA unslotted ALOHA system with successive interference cancellation, IEEE Global Telecommunications Conference (Globecom), Vol. 5, pp. 3173-3177, November 2001, San Antonio (USA). [113] Q. Liu, E.-H. Yang, Z. Zhang, Throughput analysis of CDMA systems using mul- tiuser receivers, IEEE Transactions on Communications, Vol. 49 , No. 7, pp. 1192- 1202, July 2001. [114] P. Kota, C. Schlegel, A wireless packet multiple access method exploiting joint detection, IEEE International Conference on Communications (ICC), Volume: 4, pp.2985-2989, 11-15 May 2003. [115] C. Schlegel, R. Kempter, P. Kota, A novel random wireless packet multiple access method using CDMA, IEEE Transactions on Wireless Communications, Vol. 5, No. 6, pp. 1362-1370, June 2006. [116] Karna et al., ”Data transmission method and a radio system, United States Patent No. 6625138, 23 September 2003. [117] O. del Rı́o Herrero, R. De Gaudenzi, A high efficiency scheme for quasi-real-time satellite mobile messaging systems, Proc. of the 10th International Workshop on Signal Processing for Space Communications, SPSC 2008, pp. 1-9, 6-8 October 2008, Rhodes Island (Greece). [118] O. del Rı́o Herrero, R. De Gaudenzi, A High Efficiency Scheme for Large-Scale Satellite Mobile Messaging Systems, 27th AIAA International Communications Satellite Systems Conference (ICSSC), 1-4 Jun. 2009, Edinburgh (UK). [119] K. Kim and G. Shevlyakov, Why Gaussianity?, IEEE Signal Processing Magazine, March 2008, pp. 102-113. [120] P. Patel, J. Holtzman, Analysis of a DS/CDMA successive interference cancel- lation scheme using correlations, IEEE Global Telecommunications Conference (Globecom), Vol. 1, pp. 76-80, 29 November - 2 December 1993, Houston (USA). [121] C. A. Putman, S. S. Rappaport, D. L. Schilling, A Comparison of Schemes for Coarse Acquisition of Frequency Hopped Spread Spectrum Signals, IEEE Transac- tions on Comm., February 1983.
- 99. BIBLIOGRAPHY 209 [122] G. E. Corazza, P. Salmi, A. Vanelli-Coralli, M. Villanti, Differential and non- coherent post detection integration techniques for the return link of satellite W- CDMA systems, 13th IEEE International Symposium on Personal, Indoor and Mo- bileRadio Communications, Vol. 1, pp. 300- 304, September 2002. [123] M.K. Sust, R.F. Kaufmann, F. Molitor, G.A. Bjornstrom, Rapid acquisition con- cept for voice activated CDMA communication, Global Telecommunications Con- ference (Globecom), Vol. 3, pp. 1820-1826, 2-5 December 1990, San Diego (USA). [124] M.C. Reed, Acquisition for Satellite UMTS with Large Frequency Offsets, 5th European Workshop on Mobile/Personal Satcoms (EMPS2002), 25-26 September 2002, Baveno-Stresa, Lake Maggiore (Italy). [125] 3GPP TS 25.211 v3.12.0, Physical channels and mapping of transport channels onto physical channels (FDD); Release 1999, September 2002. [126] 3GPP TS 25.213 v3.9.0, Spreading and modulation (FDD); Release 1999, January 2004. [127] ETSI TS 101 851-3 V1.1.1, Satellite Component of the UMTS/IMT2000 A-family; Part 3: Spreading and Modulation December 2000. [128] ETSI TS 101 851-4 V1.1.1 Satellite Component of UMTS-IMT 2000;A-family; Part 4: Physical layer procedures, December 2000. [129] C. E. Gilchriest, Signal-to-Noise Monitoring, JPL Space programs Summary, No. 37-27, Vol. IV, pp.169-184, 30 June 1964. [130] F. Perez-Fontan, M. Vazquez-Castro, C. Enjamio Cabado, J. Pita Garcia, E. Ku- bista, Statistical Modelling of the LMS Channel, IEEE Transactions on Vehicular Technology, Vol. 50, No. 6, November 2001. [131] F. Perez-Fontan, M. Vazquez-Castro, S. Buonomo, J.P. Poiares-Baptista, B. Arbesser-Rastburg, S-band LMS channel behaviour for different environments, de- grees of shadowing and elevation angles, IEEE Transactions On Broadcasting, Vol. 44, No. 1, pp. 40-75, March 1998. [132] ETSI TS 102 584 v1.1.1 Digital Video Broadcasting (DVB): DVB-SH Implemen- tation Guidelines, December 2008. [133] B. Wlodzimierz, The normal distribution: characterizations with applications, Springer-Verlag 1995, ISBN 0-387-97990-5. [134] W. Peter, An Introduction to Ergodic Theory, Springer, 1982. [135] G. Liva, Graph-Based Analysis and Optimization of Contention Resolution Diver- sity Slotted ALOHA, IEEE Transactions on Communications, Vol. 59, Issue 2, pp. 477-487, February 2011. [136] E. Paolini, G. Liva, M. Chiani, High Throughput Random Access via Codes on Graphs: Coded Slotted ALOHA, Proceedings IEEE International Communication Conference (ICC), 5-9 June 2011, Kyoto (Japan). [137] H.-C. Bui, J. Lacan and M.-L. Boucheret, An enhanced multiple random access scheme for satellite communications, In the Proc. of the 2012 Wireless Telecommu- nications Symposium (WTS), pp. 1-6, 18-20 April 2012, London (United Kingdom). [138] H.-C. Bui, J. Lacan and M.-L. Boucheret, Multi-slot coded ALOHA with irregular degree distribution, In the Proc. of the first International IEEE-AESS Conference in Europe on Space and Satellite Telecommunications (ESTEL), October 2012, Rome (Italy).
- 100. 210 BIBLIOGRAPHY [139] F. Ricciato, P. Castiglione, Pseudo-random Aloha for enhanced collision-recovery in RFID, IEEE Communications Letters, Vol. 17, No. 3, pp. 608-611, March 2013. [140] P. Castiglione, F. Ricciato, P. Popovski, Pseudo-Random ALOHA for Inter-frame Soft Combining in RFID Systems, In the Proc. of IEEE 2013 18th International Conference on Digital Signal Processing. 2013. [141] J. Massey, P. Mathys, The Collision Channel Without Feedback, IEEE Transac- tions on Information Theory, Vol. IT-31, No. 2, pp. 192-204, March 1985. [142] C. Kissling, Performance Enhancements for Asynchronous Random Access Pro- tocols over Satellite, Proceedings IEEE International Communication Conference (ICC), pp. 1-6, 5-9 June 2011, Kyoto (Japan). [143] F. Clazzer and C. Kissling, Enhanced Contention Resolution ALOHA - ECRA, 9th International ITG Conference on Systems, Communications and Coding (SCC), January 2013, Munich (Germany). [144] O. del Rı́o Herrero, E. Casini, R. De Gaudenzi, Contention Resolution Diversity Slotted Aloha Plus Demand Assignment (CRDSA-DA): an Enhanced MAC Protocol for Satellite Access Packet Networks, 23rd AIAA International Communications Satellite Systems Conference, 25-28 Sep. 2005, Rome (Italy). [145] ETSI TS 102 721-5 v1.1.1, Satellite Earth Stations and Systems; Air Interface for S-band Mobile Interactive Multimedia (S-MIM); Part 5: Protocol Specifications, Link Layer, December 2011. [146] Digital Video Broadcasting (DVB), Second Generation DVB Interactive Satellite System (DVB-RCS2); Guidelines for Implementation and Use of LLS: EN 301 545-2, DVB Document A162, February 2013. [147] J.O. Irwin, On the Frequency Distribution of the Means of Samples from a Popu- lation Having any Law of Frequency with Finite Moments, with Special Reference to Pearson’s Type II, Biometrika, Vol. 19, No. 3/4, pp. 225-239, 1927. [148] P. Hall, The Distribution of Means for Samples of Size N Drawn from a Population in which the Variate Takes Values Between 0 and 1, All Such Values Being Equally Probable, Biometrika, Vol. 19, No. 3/4, pp. 240-245, 1927. [149] F. Giannetti, M. Luise and R. Reggiannini, Chip timing recovery in digital modems for continuous-phase CDMA radio communications, IEEE Transactions on Com- munications, Vol. 43, Issue 234, pp. 762-766, February/March/April 1995. [150] Newtec, Sat3Play IP Broadband System, http://www.newtec.eu/product/sat3play- ip-broadband-hub. [151] Viasat, A Breakthrough in Satellite Communications: The Story of the Technology Behind ArcLight, Online Journal of Space Communication, Issue 7, VSAT Com- munication, Fall 2004. [152] European Space Agency, Advanced MODEM techniques for future satellite access networks, ESA Contract Number 4000108548/13/NL/JK, 2013. [153] F. Collard, R. De Gaudenzi, On the Optimum Packet Power Distribution for Spread Aloha Packet Detectors with Iterative Successive Interference Cancellation, IEEE Transactions on Wireless Communications, Vol. 13, No. 12, pp. 6783-6794, October 2014.
- 101. BIBLIOGRAPHY 211 [154] G. Gallinaro, N. Alagha, R. De Gaudenzi, K. Kansanen, R. Müller and P. Salvo Rossi, ME-SSA: an Advanced Random Access for the Satellite Return Channel, IEEE International Conference on Communications (ICC), 8-12 June 2015, London (UK). [155] G. Gallinaro, F. Di Cecca, M. Marchitti, R. De Gaudenzi and O. del Rı́o Herrero, Enhanced spread spectrum ALOHA system level performance assessment, Wiley In- ternational Journal of Satellite Communications and Networking, on-line published November 2013. [156] C. Kissling, A. Munari On the integration of random access and DAMA chan- nels for the return link of satellite networks, IEEE International Conference on Communications (ICC), pp. 4282-4287, 9-13 June 2013, Budapest (Hungary). [157] C. Stefanovic, K.F. Trilingsgaard, N.K. Pratas, P. Popovski, Joint estimation and contention-resolution protocol for wireless random access, IEEE International Conference on Communications (ICC), pp. 3382 - 3387, 9-13 June 2013, Budapest (Hungary). [158] F. Vazquez-Gallego, M. Rietti, J. Bas, J. Alonso-Zarate, L. Alonso, Performance Evaluation of Frame Slotted-ALOHA with Succesive Interference Cancellation in Machine-to-Machine Networks, Proceedings of 20th European Wireless Conference, pp. 1-6, 14-16 May 2014, Barcelona (Spain).
- 103. Acronyms 3G Third Generation 3GPP Third Generation Partnership Project ACRDA Asynchronous Contention Resolution Diversity ALOHA ARQ Automatic Repeat request ATM Air Traffic Management ATS Applications Technology Satellite AWGN Additive White Gaussian Noise BPSK Binary Phase Shift Keying BTMA Busy Tone Multiple Access CDMA Code Division Multiple Access CFDAMA Combined Free and Demand Assignment Multiple Access CFDAMA-PB Combined Free and Demand Assignment Multiple Access - Piggy Backed CPM Continuous Phase Modulation CPODA Contention Priority Oriented Demand Assignment CRC Cyclic Redundancy Check CRDSA Contention Resolution Diversity Slotted Aloha CRDSA-DA Contention Resolution Diversity Slotted Aloha plus Demand Assignment CRRMA Combined Random/Reservation Multiple Access CSA Coded Slotted ALOHA CSMA Carrier Sense Multiple Access CSMA/CA Carrier Sense Multiple Access / Collision Avoidance CSMA/CD Carrier Sense Multiple Access / Collision Detection CTS Clear To Send DA Demand Assignment DAMA Demand Assignment Multiple Access DEA Diploma Estudis Avançats DSA Diversity Slotted ALOHA DS-CDMA Direct Sequence - Code Division Multiple Access DS-SS Direct Sequence - Spread Spectrum DTH Direct To Home DVB Digital Video Broadcasting DVB-RCS Digital Video Broadcasting - Return Channel Satellite DVB-RCS2 Digital Video Broadcasting - Return Channel via Satellite second generation DVB-S2 Digital Video Broadcasting - Satellite DVB-S2 Digital Video Broadcasting - Satellite - Second Generation DVB-SH Digital Video Broadcasting - Satellite services to Handhelds 213
- 104. 214 ACRONYMS ECRA Enhanced Contention Resolution ALOHA EIRP Equivalent Isotropically Radiated Power ESA European Space Agency E-SSA Enhanced Spread-Spectrum ALOHA ESTEC European Space Research and Technology Centre ETSI European Telecommunications Standards Institute EUROCAE European Organisation for Civil Aviation Equipment FAMA Floor Acquisition Multiple Access FCA Free Capacity Assignment FDMA Frequency Division Multiple Access FEC Forward Error Correction FER Frame Error Rate FM Frequency Modulation GEO Geostationary Earth Orbit GMSK Gaussian Minimum Shift Keying GPS Global Positioning System HD High Definition HPA High Power Amplifier HTS High Throughput Satellite IC Interference Cancellation IEEC Institut Estudis Espacials de Catalunya IEEE Institute of Electrical and Electronics Engineers IFS Inter Frame Spacing IP Internet Protocol IPoS IP over Satellite IRSA Irregular Repetition Slotted ALOHA iSIC iterative Successive Interference Cancellation ITU International Telecommunication Union iTV interactive TV LEO Low Earth Orbit LMS Land Mobile Satellite LNB Low Noise Block HEO Highly Elliptical Orbit M2M Machine to Machine MAC Medium Access Control MACA Multiple Access Collision Avoidance MACAW Multiple Access Collision Avoidance for Wireless MAI Multiple Access Interference ME-SSA MMSE plus Enhanced Spread-Spectrum ALOHA MF-TDMA Multi-Frequency - Time Division Multiple Access MMSE Minimum Mean Square Error MMSE-SIC Minimum Mean Square Error - Successive Interference Cancellation MUD Multi User Detection MuSCA Multi-Slots Coded ALOHA NASA National Aeronautics and Space Administration NCDP Network-Coded Diversity Protocol NTSC National Television System Committee
- 105. ACRONYMS 215 PAL Phase Alternating Line PCMA Paired Carrier Multiple Access PDF Probability Density Function PLR Packet Loss Ratio PNC Physical layer Network Coding PRDAMA Predictive DAMA PSK Phase Shift Keying QPSK Quadrature Phase Shift Keying RA Random Access RACH Random Access Channel RAN RA with Notification protocol RF Radio-Frequency RFID Radio-Frequency Identification RTN Return RTS Request To Send S-ALOHA Slotted Aloha SCADA Supervisory control and data acquisition SD Standard Definition SDR Software Defined Radio SDUPTC SNIR Driven Uplink Packet Transmission Control SECAM Séquentiel Couleur à Mémoire SESAR Single European Sky ATM Research SF Spreading Factor SIC Successive Interference Cancellation S-MIM S-band Mobile Interactive Multimedia SNIR Signal to Noise plus Interference Ratio SNR Signal to Noise Ratio SSA Spread Spectrum ALOHA SSMA Spread Spectrum Multiple Access SUMF Single User Matched Filter TDM Time Division Multiplexing TDMA Time Division Multiple Access TIA Telecommunication Industry Association TT&C Telemetry, Tracking and Command TV Television UHF Ultra-High Frequency UMTS Universal Mobile Telecommunications System UW Unique Word VF Virtual Frame VSAT Very Small Aperture Terminal