Constellation Shared Multiple Access - A Noma Scheme for Increased User Capacity in 5G MMTC

International Journal of Computer Networks & Communications (IJCNC) Vol.14, No.3, May 2022
DOI: 10.5121/ijcnc.2022.14305 73
CONSTELLATION SHARED MULTIPLE
ACCESS - A NOMA SCHEME FOR INCREASED
USER CAPACITY IN 5G MMTC
Kiran V. Shanbhag1
and Savitha H. M.2
1
Department of Electronics and Communication Engineering,
Anjuman Institute of Technology and Management, Bhatkal, Karnataka, India
2
Department of Electronics and Communication Engineering,
St Joseph Engineering College, Mangaluru, India
ABSTRACT
While the legacy cyclic prefix orthogonal frequency division multiple access is retained as the preferred
multiple access scheme for 5G enhanced mobile broadband the research is now focussed on the multiple
access schemes for massive machine type communication (mMTC) and ultra-reliable low latency
communication .Though orthogonal multiple access schemes provide simple reception, they limit number
of simultaneous user equipment as against the primary requirement of mMTC. On the other hand, the
various non-orthogonal multiple access schemes which have been proposed so far as the likely solution,
need complex successive interference cancellation receivers. So a simplified scheme named constellation
shared multiple access is proposed here which substantially increases the number of simultaneous users to
be served within a single resource block (RB) in LTE or 5G New Radio, thus aiding the massive
connectivity requirement of mMTC. This is achieved by differentiating among the users in constellation
domain. Moreover, the simple architecture compatible with 5G eMBB makes it a strong contender multiple
access contender for 5G mMTC.
KEYWORDS
Non orthogonal multiple access, Quality of service, Massive connectivity, 5G, Successive interference
cancellation receiver, mMTC.
1. INTRODUCTION
Fifth generation (5G) wireless network is now almost a commercial reality as the standardization
is under process. 5G networks support three major applications, which include enhanced mobile
broadband (eMBB), massive machine- type communications (mMTC) and ultra-reliable and low-
latency communications (URLLC). These scenarios require massive connectivity, high system
throughput and improved spectral efficiency (SE) as compared to long term evolution (LTE) [1].
In order to meet these new requirements, new modulation and multiple access (MA) schemes are
being explored. Though CP OFDMA has been almost adapted as the preferred technique for
eMBB due to its ability to enhance communication speed, the synchronized nature of the scheme
limits the number of User Equipment (UE) that can be simultaneously supported. But one of the
key requirements of mMTC and URLLC is the need to support a large number of UEs that not
only try to communicate simultaneously but with minimum involved overhead which are the
issues that OFDMA cannot address as is. As a result several Non-Orthogonal Multiple Access
(NOMA) schemes are being proposed as a solution. The unsynchronized manner in which the
schemes work, enables them to not only serve non uniform number of UEs simultaneously but
also to provide grantless access which reduces overheads thus improving the latency [2],[3].

74
In a typical NOMA system, signal transmitter and receiver are jointly optimized, so that multiple
layers of data from more than one UE can be simultaneously delivered using the same time,
frequency and spatial resource in a superimposed manner. At the receiver side, the information of
different UEs can be recovered by advanced receivers such as successive interference
cancellation (SIC) or iterative detection receivers [4]. This differs from conventional power
allocation strategies, such as water filling, as NOMA allocates less power for the users with better
downlink channel state information (CSI), to guarantee overall fairness and to utilize diversity in
the time /frequency/ code domains. The user with more transmit power is first to be decoded
while treating the other user’s signal as noise. Once the signal corresponding to the user with the
larger transmit power is detected and decoded, its signal component will be subtracted from the
received signal by SIC receiver to facilitate the detection of subsequent users. The difference
among these schemes is mainly on UE’s signature design, i.e., whether the scrambling sequence,
interleaving or spreading code is used to differentiate UEs.
1.1. Related Work
One of the earliest NOMA schemes proposed is a scheme called sparse code multiple access
(SCMA), a CDMA based scheme but with sparse codes and multidimensional constellation to
separate simultaneous users in same resource block [5]. But the scheme is prone to user collision
needing complex reception under overloaded scenario. Moreover the scheme, like any other
multicarrier modulation (MCM) schemes, is suffering from high PAPR and there is no mention
of single carrier version as solution. Another prominent scheme presented is named resource
spread multiple access (RSMA) which uses low rate channel codes and scrambling with good
correlation properties to separate users , also needing SIC receivers [6]. Interleave Division
Multiple Access another scheme which uses the concept of interleaving to increase user capacity
[7]. Though all these schemes tend to enhance simultaneous user capacity, it should be noted that
the superimposed nature of data forces complex iterative reception which may not be efficient in
terms of power consumption, receiver hardware complexity as against the requirements of
mMTC. A detailed survey of several such NOMA contenders classify them as either scrambling
based, spreading based or interleaving based but with SIC based receiver as a common challenge
[8]. It’s also observed that as more and more users are accommodated, either the performance
deteriorates drastically at receiver side or needs complex iterative reception. The recent study has
[9] also revealed that 5G systems are likely to operate with higher order modulation schemes
such as 1024 QAM, 2048 QAM and 4096 QAM in millimetre wave frequencies without much
bandwidth constraint but with high connection density requirements of about 1M devices/Km2.
The coexistence of eMBB with mMTC and uRLLC is also a research concern with compatibility
being the preference. The 5g services in general are expected provide a wide range of quality of
service (QoS) ie provide different priority among different applications in terms of bitrate,
latency, jitter etc or to guarantee a certain performance in terms of data rate/ throughput . This is
usually done by having a suitable QoS flow identifier (QFI)[10], [11].
In order to address these issues, a new multiple access scheme named constellation shared
multiple access (CSMA) is presented here. It uses non overlapping constellation points allocation
to differentiate among several simultaneous users within a set of subcarriers in a LTE resource
block. This not only enhances the number of users, but also the non-overlapping nature of the
resource allocation i.e. constellation points in this case, greatly simplifies the reception process as
against SIC reception [12][13]. The scheme can also satisfy wide QoS requirements that arise in
mMTC scenarios. The CSMA scheme can provide for the bitrate/ throughput aspects of QoS
flow by supporting different guaranteed bit rate (GBR) and non GBR flows as it can flexibly
schedule the constellation resources to achieve minimum requirements to certain users.The
scheme proposed is a simple extension of legacy CP OFDMA in LTE such that, it allows the
existing set of subcarriers in a single resource block(RB) that would have been otherwise used by

75
a single UE in OFDMA block, to be now utilized by multiple UE by separating them in the
constellation domain as compared scrambling, interleaving or code domain as suggested in other
NOMA approaches. This results in a substantial increase in the number of UEs that can be
simultaneously serviced. The scheme also has advantage over other NOMA schemes due to the
following reasons. Firstly, it’s the similarity and compatibility with OFDMA, which aides in
simple migration to NOMA mode when more UEs are present. Secondly, the non-overlapping
nature of constellation mapping, which ensures simplicity of the decoding scheme, as compared
to the SIC reception in case of other NOMA schemes. Throughout this study, we will assume an
underlying OFDMA layer that takes care of synchronization, inter-symbol interference (ISI) etc
and the focus will be mainly on accommodating/differentiating users in a single RB built on this
OFDMA layer. Section 2 illustrates the proposed scheme, discusses in detail the improvement in
user capacity, the various advantages and disadvantages of the scheme. Section 3 includes the
simulation results and section 4 includes the conclusion and scope for further research on the
scheme.
2. PROPOSED SCHEME
The proposed scheme is developed as an extension of existing CP OFDMA based LTE
architecture. For orthogonal multiple access like OFDMA, as per the standard terminology, a RB
is the smallest unit of resource that can be allocated to a single user[14]. For an FDD (frequency
domain duplex) based frame structure, it comprises of 12 subcarriers in frequency domain along
with one time slot comprising of either 7 OFDM symbols in LTE or 14 OFDM symbols in 5G
NR[15]. While the data from a single UE is usually QAM modulated to be mapped onto an entire
subcarrier resource block, the idea in proposed scheme is to accommodate multiple users that are
separated in QAM constellation, in the same RB in a time shared manner. This is done by
mapping the data bits of individual UE to a higher constellation by using either a look up table or
suitable mapping logic.
Figure 1. A sample lookup table based higher constellation mapping scheme where four UE data with 4
QAM constellation is being mapped to a single 16 QAM constellation
An example scheme is depicted in figure 1, where each 2 bit symbol data from 4 users, which
would need a QAM modulation with 4 constellation points, have been mapped to 4 bits such that
the resulting 16 constellation points are equally divided among 4 users, as per the look up table 1.
This should allow a single RB with 16 QAM modulated OFDM symbols, to serve four.
Higher
QAM
Constellation
Mapping
Logic
00
10 11
01
0000
0001
0100 1100 1000
0101 1101 1001
0011 0111 1111 1011
0010 0110 1110 1010
UE1
UE2
UE3
UE4
00
10 11
01
00
10 11
01
00
10 11
01

76
Table 1. Look up table for the scheme in figure 1, where each 2 bit symbol data of 4 UEs are being mapped
to unique 4 bit symbol data of 16 QAM constellation
User User Data
Corresponding Higher
Constellation Point
UE1
00
01
10
11
0100
1000
0111
1011
UE2
00
01
10
11
0000
1100
0011
1111
UE3
00
01
10
11
0101
1001
0110
1010
UE4
00
01
10
11
0001
1101
0010
1110
UE, that are separated in constellation space. To avoid additional signalling overhead per user, it
is assumed that all the users sharing the RB experience similar channel conditions. Another
approach for mapping onto the higher constellation would be to simply combine unique user
identifier (address) bits along with their data bits in a particular bit location so as to share QAM
constellation points equally among several UEs. Say, if the total available constellation points at
any instant in M-QAM such that M= 2D
, then these are to be shared among several UEs in a
systematic manner such that fixed ‘A’ bits out of ‘D’ are used for addressing 2A user equipment
and remaining bits are used for actual data . If, for example, 64 QAM is being used, and the
resources are to be shared among 4 users, then each UE gets 16 unique constellation points,
amounting to 4 data bits per user. Then out of six bits which would form 26
i.e. 64 constellation
point, any 2 bits are used as part of address of UE and remaining 4 bits are allocated for actual
data. Similarly, if any 3 bits out of 6 bits are used to identify UEs, the scheme can increase
simultaneous user capacity 8 times and so on. The table 2 depicts one of the schemes for 64
QAM where, out of 6 bits, the central 2 bits represent one of the 4 UE’s addresses, whereas first
and last 2 bits represent the actual data of corresponding user. Figure 2 shows the corresponding
constellation diagram where the UEs are sharing all 64 constellation.
Table 2. An example scheme where out of 6 bits constituting a constellation point, 2 bits in centre indicate
UE address and remaining bits indicate the actual UE data
points equally among each other. Thus the 4 UEs can now share the same RB in a time shared
manner, though with reduced throughput. As indicated by the diagram, the constellation points
D5 D4 D3 D2 D1 D0 D5 D4 D3 D2 D1 D0 D5 D4 D3 D2 D1 D0 D5 D4 D3 D2 D1 D0
B3 B2 A1 A0 B1 B0 B3 B2 A1 A0 B1 B0 B3 B2 A1 A0 B1 B0 B3 B2 A1 A0 B1 B0
0 0 0 0 0 0
0 0 0 0 0 1
x x 0 0 x x
x x 0 0 x x
1 1 0 0 1 0
1 1 0 0 1 1
0 0 0 1 0 0
0 0 0 1 0 1
x x 0 1 x x
x x 0 1 x x
1 1 0 1 1 0
1 1 0 1 1 1
0 0 1 0 0 0
0 0 1 0 0 1
x x 1 0 x x
x x 1 0 x x
1 1 1 0 1 0
1 1 1 0 1 1
0 0 1 1 0 0
0 0 1 1 0 1
x x 1 1 x x
x x 1 1 x x
1 1 1 1 1 0
1 1 1 1 1 1
UE 1 data
combined with
address bits 0 0
UE 2 data
combined with
address bits 0 1
UE 3 data
combined with
address bits 1 0
UE 4 data
combined with
address bits 1 1

77
corresponding to individual UE are clearly differentiated from each other and do not overlap.
Both the mapping schemes, the one with the look up table and the illustration with 2 centre bits
used as user address are straight forward simple schemes among several possibilities, meant for
the readers understanding, but may not be an ideal one. The optimum allocation and the
distribution of user constellation points should be done by taking into consideration, the shaping
gain as well as PAPR experienced by each UE.[16], [17]
Figure 2. The CSMA scheme as per table 2, where 64 QAM constellation points are shared by 4 UEs as per
legend
2.1. Features and limitations
The following section describes the advantageous features of the scheme such as user capacity
expansion, provision of QoS and also the impact of the scheme on the throughput.
2.1.1. User Capacity Expansion
Depending on the existing channel conditions and available SNR, which in turn determine the
kind of modulation scheme to be used, the numbers of concurrent users supported also vary.
Table 3 summarizes likely increase in capacity (Uc) for a particular modulation scheme, based on
‘B’, the symbol width of actual data in individual UEs and ‘M’ the higher constellation scheme
employed for multiplexed data , given by the expression
Uc = 2(log2M−B)
(1)
Table 3. The User capacity enhancement factor for various modulation schemes and different symbol
widths of the UE
Modulation
Scheme
M
Symbol width of
UE
‘B’ bits
User capacity
Enhancement
Factor , Uc
64 QAM 4 4
64 QAM 3 8
64 QAM 2 16
256 QAM 6 4
256 QAM 4 16
256 QAM 2 64
1024 QAM 8 4
1024 QAM 6 16
1024 QAM 4 64

78
2.1.2. Provision of Quality of Service (QoS)
The scheme can also be used to provide different QoS among shared UEs by doing non uniform
constellation scheduling similar to guaranteed bit rate (GBR) bearer. There are two means of
doing so. One is by having a look up table with more points allocated to particular UE along with
allocating relatively more OFDM symbols per slot per RB for such UE with higher throughput
requirements. Figure 3 shows one such scheme where where out of 16 constellation points, UE1
has been allocated 8 , thus being able to send 3 information bits per symbol while UE2, UE3 get
4 points each and are able to send only 2 bits . Table 4 shows the corresponding look up table.
Secondly, the bit allocation for address and data can be changed dynamically as per the QoS
requirements for a certain group of UEs, which can be together multiplexed to a common QAM
modulator block. This way, while few of 64 QAM modulators can serve say, 4 UE by using 4 bit
symbol width for better throughput, remaining can serve 8 UE by using 3 bit symbol width for
with reduced throughput.
0000
0001
0100 1100 1000
0101 1101 1001
0011 0111 1111 1011
0010 0110 1110 1010
0000
0001
0100 1100 1000
0101 1101 1001
0011 0111 1111 1011
0010 0110 1110 1010
UE1
UE2
UE3
UE4
(a) (b)
Figure 3. A comparison of uniform constellation sharing among users in (a) vs non uniform constellation
sharing in (b) where UE1 has more priority
Table 4. Look up table where UE1 has high throughput compared to UE2, UE3
User
User
Data
Corresponding Higher
Constellation Point
UE1
000
001
010
011
100
101
110
111
0100
1000
0111
1011
0001
1101
0010
1110
UE2
00
01
10
11
0000
1100
0011
1111
UE3
00
01
10
11
0101
1001
0110
1010

79
2.1.3. Reduced receiver complexity
Most of NOMA schemes proposed as of now, rely on successive interference cancellation (SIC)
receivers which are fairly complex, iterative in nature and are likely to be less power efficient as
they need knowledge of channel state information for differential resource allocation,
interconnection among all UE for subsequent reception, thus affecting the concern of long battery
life in mMTC devices [18]. The proposed scheme, even though requires entire subcarrier data to
be processed by all UE, since they are well separated in constellation space, does not need of SIC
reception. The receiver complexity would be that of conventional OFDMA receiver itself, as per
our earlier assumption along with an additional reference to the lookup table after QAM
demodulation by each UE to receive corresponding data.
2.1.4. Compatibility with existing platforms and future technologies
While most of the NOMA schemes suggested so far go for mostly clean slate approach as
compared to 5G eMBB, needing additional infrastructure for those applications which need high
UE support, the proposed scheme can be easily adapted with slight modifications to legacy
OFDMA infrastructure. Depending on the scenario, the switch between limited user, bandwidth
extensive applications and the massive connectivity low bit rate applications can be easily
achieved using suitable control signals. As depicted in fig 1, fig 3, all of the QAM constellation
points, along with a RB can be either allocated to single user in an orthogonal manner giving high
bit rate or it can be shared in time/constellation space to accommodate more UE. With the
introduction of OFDMA in WiFi 6, which is another platform trying to capture the IoT space
[19][20], the suggested scheme can be easily extended to address massive connectivity issues.
2.1.5. Impact on throughput and bit error rate (BER)
The increased user capacity comes at the cost of reduced data rate as fewer bits are now available
for the actual data due to addressing overhead, compared to CP OFDMA case and each shared
user is transmitting in a multiplexed manner. But as the target mMTC applications are more
focussed towards massive UE connectivity and not necessarily on high bandwidth applications as
compared to eMBB, the reduced bit rate is not of much concern as the likely bit rates in mMTC
can go as low as few hundred bits per second[21].
The throughput reduction is mainly due to
1. The addressing overhead of ‘A’ bits out of log2M available bits for each UE
identification reduces bit rate by factor (log2M-A)/ log2M , which further reduces with
increasing users, though not in an exponential manner.
2. The multiplexed nature of user data transmission, which forces only a single user among
2A UE connected to common modulator, to transmit/receive data at any instant of time.
Together for an M QAM scheme, out of ‘D’ bits, for ‘A’ number of bits used for addressing, the
total reduction in throughput can be given by
Tr = (log2M − A) ((log2M) ∗ 2A)
⁄ (2)
The BER performance of the scheme is relatively low as compared to its orthogonal counterpart
where a single UE is connected to a dedicated M QAM modulator. This is due to the fact that out
of ‘D’ bits forming an M QAM symbol, only ‘B’ bits carry actual data. The addition of more and
more addressing bits to accommodate more UEs, forces the use of higher order modulation,
which usually have poorer performance, due to the proximity among the constellation points

80
pertaining to other UE.The probability of symbol error (Ps) or the symbol error rate (SER) for
any M QAM modulation scheme is given as under
Ρs = 1 − (1 −
2√M−1
√M
Q [√
3υs
M−1
])
2
(3)
Here ϑs is the SNR per symbol given as ϑs =
εs
η0
and εs is energy per symbol and η0 is the
noise power. It is also related to SNR per bit given by ϑs = ϑblog2M where ϑb =
εb
η0
and εb is
energy per bit. Say, if for same throughput from a single user perspective, if B bits are allocated
for data, then M=2B
QAM is used and the expression for Ps in terms of symbol width B is given
as
Ρs = 1 − (1 −
2√2B−1
√2B
Q [√
3υs
2B−1
])
2
(4)
However, if ‘A’ additional bits are appended to accommodate more users, 2(B+A)
QAM is used,
which will have relatively poor performance due to denser constellation point arrangement given
in terms of symbol and address bit width by Ps’
Ρs′ = 1 − (1 −
2√2B+A−1
√2B+A
Q [√
3υs
2B+A−1
])
2
(5)
2.2. Comparison with other NOMA schemes
Most of existing NOMA schemes can be classified as either code domain NOMA (CD-NOMA)
or power domain NOMA (PD-NOMA). The intention of most of these schemes is to provide
massive connectivity and low latency via grantless access. As compared to the proposed scheme,
the following observations are to be made with respect to existing NOMA schemes.
2.2.1. Relative latency and power consumption
The grantless access in most NOMA schemes is owing to the fact that no access grant signalling
is needed by each UE. But to be able to decode the data pertaining to the last user, all the
previous data pertaining to the earlier users shall be recovered iteratively. This SIC reception will
in turn reduce battery life which worsens with increasing users. In the proposed scheme however,
once the number of users are determined, a RB is granted in time shared manner, without the
need for any SIC reception. This scheme, even though slows down the transmission, ensures that
a fixed number of users per RB are served, with uniform resources reserved per users, providing
almost uniform BER performance, requiring simple reception/transmission and uniform
processing time.
2.2.2. Impact of higher UE on performance
Figure 4 shows proposed scheme with 4 users sharing a 16 qam constellation with 4, 4 QAM
constellation points each. It clearly shows that per user performance will drop to that of a 16
QAM per user, even though its assigned only 4 constellation points. An important point here is
to note that for CSMA constellations pertaining to 4 users sharing single RB, even though as a
whole equals that of a 16 QAM, since at any instant within a particular time slots, for a particular

81
user there can only be 4 constellation possibilities as in 4 QAM making it relatively robust.
Compare it to figure 5 which shows a mere 3 user PD NOMA constellation resembles
DE16
t1
t2
t3
t4
UE1
UE2
UE3
UE4
DE4
Figure 4. An example 16 QAM constellation being shared by 4 UE, each with 4 points (left) , in time
multiplexed manner with t1,t2,t3,t4 representing different intervals within a time slot where they transmit,
with each user constellation points enjoying uniform distance from each other (right)
almost a 64 QAM, moreover requiring SIC reception. While most literatures pertaining to
NOMA illustrate 2 user cases, higher user numbers reveal another downside. Here, each user data
is supposed to be 4 QAM modulated and superimposed in power domain. Those users who do
the initial QAM decoding are much reliable as the minimum Euclidian distance, DE4 among
corresponding constellation is much higher as shown in figure 5.b. But for the subsequent users,
for same bit rate, the constellation points get much closer and closer, affecting the BER
performance. As seen from fig 5.c. for UE2, the constellation resembles to that of an inferior 16
QAM with distance DE16. Again from fig 5.d the constellation for UE3 is much inferior
resembling that of 64 QAM with distance DE64 thus relying heavily on the error control codes
for proper decoding. Having 4th user will make it much worse. This decay with increasing users
is much faster in PD- NOMA compared to proposed scheme , as a 4 user scheme CS NOMA is
doing much better compared to a mere 3 user PD NOMA, that too without needing SIC.
DE16
DE64
DE4
UE1
UE2
UE3
Figure 5. An example 3 user PD NOMA constellation space (top left) with each user employing 4 QAM,
with illustrations of user 1 space, along with relatively low power user 2 signal on top (top right), user 2
constellation space, along with low power user 3signal on top of it (bottom left) clearly depicting Euclidean
distance reduction with increasing users

82
2.2.3. Ease of scalability
While in most NOMA, accommodating more number of users involve either a robust CSI
feedback mechanism in case of PD NOMA [22] or an elaborate codebook design in case of
SCMA [23][24] and so on, for the proposed scheme however, the addition of users happens in an
exponential manner with relative ease as per a look up table or mechanism discussed in section
2.1.1, with provision to even accommodate various QoS with ease.
3. METHOD
The downlink transmitter block diagram of the proposed multiple access schemes is as shown in
figure 6. It’s a simple extension of the conventional CP OFDMA transmitter architecture for
LTE except for the initial multiplexing stage represented by the dashed area. In this initial stage,
the actual data bits corresponding to maximum of N simultaneous users such that N=2A are
mapped to a higher QAM constellation either by suitable constellation look up table (LUT)
mapping or combined with ‘A’ address bits and fed to one of the K QAM modulators
multiplexed in a time shared manner, as against LTE where each modulator would be used for
only single user. Thus, the block which would earlier accommodate K users can now
accommodate K*N number of users.
Figure 6. A typical set up of the CSMA scheme for the downlink, which is very similar to CP OFDMA,
based LTE Transmitter from the QAM modulator onwards
This higher constellation mapped data is fed to the QAM modulator either based on priority or
uniform time shared manner. Here on, the combined user data will be treated as if it were a single
UE data and transmitted as OFDMA signal. At the receiver, a reverse process is carried out,
where all the data corresponding to a particular demodulator is decoded by all connected users
and separated among several UEs once the QAM demodulation is carried out, by identifying the
location of the constellation points. Though separated in constellation space, since the same set of
subcarriers is shared by several UEs, the scheme is no more an orthogonal scheme, but a NOMA
scheme. While 5g eMBB employs single carrier version SCFDMA for uplink transmission to
address the peak to average power (PAPR) issue [25],[26], CSMA is also easily adaptable into
the single carrier version with little modifications for the uplink transmission. These features
make it a strong contender for the 5G, along with CP OFDMA. While usually in LTE or even in
5G NR, a RB is the smallest resource that can be allocated to a single user as in [27],[28],[29]
this new approach can accommodate more users per RB by allocating the OFDM symbols within
a slot among a group of users, thus increasing the user capacity. Figure 7 shows a sample 5G NR
frame structure where a RB in slot 3 containing 14 symbols per slot, is being shared among users
indicated by different colours. The users can either alternatively send their OFDM symbols in
Sub-
carrier
Mapper
L point
IFFT P/S Add
CP
S/P
S/P
S/P
QAM
Mod 1
QAM
Mod 2
QAM
Mod K
UE1
UE2
UEN
Const.
LUT
Const.
LUT
Const.
LUT

83
different time instants within a single slot or in any order as they can be identified by their
address bits i.e. constellation location.
Figure 7. An example 5G frame structure depicting the means of sharing a single RB among several users
with each being assigned alternate OFDM symbols within a slot
Number of concurrent users per RB and the possible increase in the number of users is limited by
either the order of the highest QAM constellation used based on SNR availability or the number
of OFDM symbols per slot in a resource block, if at all the existing 5G NR frame structure
architecture is to be used.
4. SIMULATION RESULTS
Figure 8 depicts several of the constellation mapping schemes for 64 QAM where out of 6 bits, 2
bits are used for addressing. Matlab software is the tool used for the simulation purpose. While
figures 8(a), 8(b) show the constellation sharing obtained by appending 2 address bits to 4 data
bits and prepending 2 address bits to data bits respectively, figures 8(c) is obtained by suitable
lookup table mapping and 8(d) is obtained by using first and fourth bits as addressing bits. It’s
observed that while few schemes distribute points uniformly over entire space, few have
constellation points in separate quadrants and so on. However, the look up table based approach,
allows much more flexibility in constellation allocation. The constellation point mapping among
various UE under a single modulator can be devised, taking into consideration several
performance parameters such as required quality of service (QoS), Euclidean distance for better
differentiation among UE, the spectral efficiency due to phase transitions, constellation shaping
gain etc. It’s observed that the users can be separated easily and each UE receivers are supposed
to receive data corresponding to relevant points after demodulating entire data. A better
constellation mapping would be the one which will have all the points distributed with maximum
distance, for same user and distributed in entire space so that even if all shared users are not
present, the data can be better differentiated from each other with better BER[30].
Figure 9 shows the simulation results of performance comparison of individual UE sharing the
RB with three more users with that of a single UE using entire RB as per the CSMA scheme,
which clearly shows the faithful recovery of sent data but at the cost of additional SNR
One Radio Frame=10 ms
One subframe = 1 ms
One slot
=0.5 ms
…………….
0 1 2 3 18 19
0 1 2 3 18 19
1 Resource
block with 12
subcarriers
and 14
OFDM
symbols
Example Resource block
shared by several users
dividing 14 OFDM
symbols among
themselves, separated in
constellation
UE 1
UE 2
UE 3
UE 4

84
requirement. Normalized BER was used instead of SER for fair comparison. 256 point FFT was
used for the OFDM with AWGN. No pulse shaping, equalization and error control coding have
been employed during simulation. For the conventional OFDMA case with single user per RB ,
16 QAM was used and assuming 4 UE sharing same modulator, requiring 2 additional bits along
with 4 data bits, 64 QAM was used with 6 bit symbol Hence, while the former performed better
due to lower modulation order, the latter experienced a denser constellation space with reduced
Euclidean distance among points as shown in figure 10, due to the increased modulation order,
as evident in the BER curve. Though there can be several factors to be taken into consideration
in choosing suitable mapping, most of them perform almost similar in terms of BER performance
as all receivers are required to demodulate under highest QAM order. This is depicted in fig
(a) (b)
(c) (d)
Figure 8. Several possible means of 64 QAM constellation sharing among 4 users
(a) Last 2 bits out of 6 are address bits (b) First 2 bits out of 6 are address bits
(c) Equidistant mapping based on LUT (d) First and 4th are address bits

85
Figure 9. BER performance comparison of the proposed CSMA scheme with conventional OFDMA with
single user allocation per RB
As explained in earlier section, it is seen that BER performance deteriorates gradually with
increasing number of users that share the constellation space per resource block. Figure 10 shows
the impact of
Figure 10. Constellation comparison of single user allocation per RB (left) with same RB shared by 4 users
(left)
accommodating more number of users in a constellation validating the equation 5. The reason for
this is further revealed in the constellation plots given in figure 11 simulated for a normalized but
relatively lower SNR of 20 dB. While the single user per RB has its constellation points much
apart from each other, those with 4 users per RB are likely to suffer [31]. While the allocation of
an RB to single user with 64 QAM modulation has resulted in good BER performance, a relative

86
decay in BER performance is observed as the number of users sharing the RB increase as
observed in a 4 and 8 users per RB respectively, resulting in higher SNR requirements.
Figure 11. BER performance comparison of the scheme for single user, 4 user and 8 user sharing single RB
employing 64 QAM OFDM
5. CONCLUSION
This paper presented a NOMA scheme to increase the simultaneous user capacity, especially for
mMTC scenario. The scheme worked by sharing a single RB among multiple users by separating
them in constellation domain, which was a slight modification to OFDMA based 5G NR
approach where the entire block was allocated to single user only. The means of separating the
users in constellation domain and the means of sharing the symbols in RB was depicted.
Analysis on the increase in user capacity, its impact on throughput and the BER was conducted.
Simulation results showed the various constellation allocation and performance comparison of
the scheme with conventional transmission. The scheme is capable providing the much required
increase in user capacity with slight modification in existing LTE or 5G NR structure. The
simplicity with which the users share the same RB and the trivial means of differentiating their
corresponding data after reception never necessitates the need for complex receivers. These
features make the CSMA scheme a strong contender for the 5G mMTC. The possibilities to
provide different QoS, optimum constellation mapping to achieve shaping gain and better PAPR
need to be explored in future.
CONFLICTS OF INTEREST
The authors declare no conflict of interest.

87
REFERENCES
[1] J. Navarro-Ortiz, P. Romero-Diaz, S. Sendra, P. Ameigeiras, J. J. Ramos-Munoz and J. M. Lopez-
Soler, "A Survey on 5G Usage Scenarios and Traffic Models," in IEEE Communications Surveys &
Tutorials, vol. 22, no. 2, pp. 905-929, Secondquarter 2020, doi: 10.1109/COMST.2020.2971781.
[2] L. Dai, B. Wang, Z. Ding, Z. Wang, S. Chen and L. Hanzo, "A Survey of Non-Orthogonal Multiple
Access for 5G," in IEEE Communications Surveys & Tutorials, vol. 20, no. 3, pp. 2294-2323,
thirdquarter 2018, doi: 10.1109/COMST.2018.2835558.
[3] Y. Yuan et al., "Non-orthogonal transmission technology in LTE evolution," in IEEE
Communications Magazine, vol. 54, no. 7, pp. 68-74, July 2016, doi: 10.1109/MCOM.2016.7509381.
[4] Souvik Sen et al., “Successive interference cancellation: a back-of-the-envelope perspective”. in
Proceedings of the 9th ACM SIGCOMM Workshop on Hot Topics in Networks (Hotnets-IX).
Association for Computing Machinery, New York, NY, USA, Article 17, 1–6. DOI:
/10.1145/1868447.1868464
[5] Y. Wu, C. Wang, Y. Chen and A. Bayesteh, "Sparse Code Multiple Access for 5G Radio
Transmission," 2017 IEEE 86th Vehicular Technology Conference (VTC-Fall), Toronto, ON, 2017,
pp. 1-6, doi: 10.1109/VTCFall.2017.8288106.
[6] Y. Cao, H. Sun, J. Soriaga and T. Ji, "Resource Spread Multiple Access - A Novel Transmission
Scheme for 5G Uplink," 2017 IEEE 86th Vehicular Technology Conference (VTC-Fall), 2017, pp. 1-
5, doi: 10.1109/VTCFall.2017.8288412.
[7] Yang Hu and Li Ping, “13. Interleave Division Multiple Access (IDMA)” in Multiple Access
Techniques for 5G Wireless Networks and Beyond , vol. 4, Dr. Mojtaba Vaezi, Prof. Zhiguo Ding,
Prof. H. Vincent Poor, Springer International Publishing , New York City, USA,2019, chapter 13, pp.
417–449.
[8] Yang Shan, Chen Peng, Liang Lin, Zhu Jianchi, And She Xiaoming, “Uplink Multiple Access
Schemes for 5G: A Survey” in ZTE Communications, Hefei, China, June 2017,vol. 15 No. S1, pp.
31–40.
[9] A. Dogra, R. K. Jha and S. Jain, "A Survey on Beyond 5G Network With the Advent of 6G:
Architecture and Emerging Technologies," in IEEE Access, vol. 9, pp. 67512-67547, 2021, doi:
10.1109/ACCESS.2020.3031234.]
[10] F. Capozzi, G. Piro, L. A. Grieco, G. Boggia and P. Camarda, "Downlink Packet Scheduling in LTE
Cellular Networks: Key Design Issues and a Survey," in IEEE Communications Surveys & Tutorials,
vol. 15, no. 2, pp. 678-700, Second Quarter 2013, doi: 10.1109/SURV.2012.060912.00100.
[11] M. Dighriri, A. S. D. Alfoudi, G. M. Lee, T. Baker and R. Pereira, "Comparison Data Traffic
Scheduling Techniques for Classifying QoS over 5G Mobile Networks", 2017 31st International
Conference on Advanced Information Networking and Applications Workshops (WAINA), Mar. 2017
[12] S. M. R. Islam, N. Avazov, O. A. Dobre and K. Kwak, "Power-Domain Non-Orthogonal Multiple
Access (NOMA) in 5G Systems: Potentials and Challenges," in IEEE Communications Surveys &
Tutorials, vol. 19, no. 2, pp. 721-742, Secondquarter 2017, doi: 10.1109/COMST.2016.2621116.
[13] Q. Luo et al., "An Error Rate Comparison of Power Domain Non-Orthogonal Multiple Access and
Sparse Code Multiple Access," in IEEE Open Journal of the Communications Society, vol. 2, pp.
500-511, 2021, doi: 10.1109/OJCOMS.2021.3064504.
[14] Arunabha Ghosh, Jan Zhang, Jefferey Andrews, Riaz Mohammed, Fundamentals of LTE, Prentice
Hall, Communications Engg. and Emerging Technologies. 2011
[15] X. Lin et al., "5G New Radio: Unveiling the Essentials of the Next Generation Wireless Access
Technology," IEEE Communications Standards Magazine, vol. 3, no. 3, pp. 30-37, September 2019,
doi: 10.1109/MCOMSTD.001.1800036.
[16] Vasudevan, Kasturi. (2008). Digital Communications and Signal Processing,Universities
Press,January 2008, chapter 3, pp.143-159
[17] Fuentes, Manuel, Louis Christodoulou and Belkacem Mouhouche. “Non-Uniform Constellations for
Broadcast and Multicast in 5G New Radio.” 2018 IEEE International Symposium on Broadband
Multimedia Systems and Broadcasting (BMSB) (2018): 1-5.
[18] Mahmoud Aldababsa, Mesut Toka, Selahattin Gökçeli, Güneş Karabulut Kurt, Oğuz Kucur, "A
Tutorial on Nonorthogonal Multiple Access for 5G and Beyond", Wireless Communications and
Mobile Computing, vol. 2018, Article ID 9713450, 24 pages, 2018. doi:10.1155/2018/9713450

88
[19] E. Khorov, A. Kureev, I. Levitsky and I. F. Akyildiz, "Prototyping and Experimental Study of Non-
Orthogonal Multiple Access in Wi-Fi Networks," in IEEE Network, vol. 34, no. 4, pp. 210-217,
July/August 2020, doi: 10.1109/MNET.011.1900498.
[20] Q. Wang, R. Zhang, L. -L. Yang and L. Hanzo, "Non-Orthogonal Multiple Access: A Unified
Perspective," in IEEE Wireless Communications, vol. 25, no. 2, pp. 10-16, April 2018, doi:
10.1109/MWC.2018.1700070.
[21] K. I. Pedersen, G. Berardinelli, F. Frederiksen, P. Mogensen and A. Szufarska, "A flexible 5G frame
structure design for frequency-division duplex cases," in IEEE Communications Magazine, vol. 54,
no. 3, pp. 53-59, March 2016, doi: 10.1109/MCOM.2016.7432148.
[22] Kryukov, Y.V., Pokamestov, D.A., Rogozhnikov, E.V. (2019). “PD-NOMA Power Coefficients
Calculation While Using QAM Signals”. In: Vishnevskiy, V., Samouylov, K., Kozyrev, D. (eds)
Distributed Computer and Communication Networks. DCCN 2019. Communications in Computer
and Information Science, vol 1141. Springer, Cham. https://doi.org/10.1007/978-3-030-36625-4_13
[23] M. Taherzadeh, H. Nikopour, A. Bayesteh and H. Baligh, "SCMA Codebook Design," 2014 IEEE
80th Vehicular Technology Conference (VTC2014-Fall), 2014, pp. 1-5, doi:
10.1109/VTCFall.2014.6966170.
[24] M. Vameghestahbanati, I. D. Marsland, R. H. Gohary and H. Yanikomeroglu, "Multidimensional
Constellations for Uplink SCMA Systems—A Comparative Study," in IEEE Communications
Surveys & Tutorials, vol. 21, no. 3, pp. 2169-2194, thirdquarter 2019, doi:
10.1109/COMST.2019.2910569.
[25] Joydeep Acharya; Long Gao; Sudhanshu Gaur, "LTE Signal Structure and Physical Channels,"
in Heterogeneous Networks in LTE-Advanced , Wiley, 2014, pp.45-71, doi:
10.1002/9781118693964.ch3.
[26] M. Chafii, F. Bader and J. Palicot, "SC-FDMA with index modulation for M2M and IoT uplink
applications," 2018 IEEE Wireless Communications and Networking Conference (WCNC), 2018, pp.
1-5, doi: 10.1109/WCNC.2018.8377028.
[27] A. Hammoodi, L. Audah and M. A. Taher, "Green Coexistence for 5G Waveform Candidates: A
Review," in IEEE Access, vol. 7, pp. 10103-10126, 2019, doi: 10.1109/ACCESS.2019.2891312.
[28] T. Salam, W. U. Rehman and X. Tao, "Data Aggregation in Massive Machine Type Communication:
Challenges and Solutions," in IEEE Access, vol. 7, pp. 41921-41946, 2019, doi:
10.1109/ACCESS.2019.2906880.
[29] I. B. F. de Almeida, L. L. Mendes, J. J. P. C. Rodrigues and M. A. A. da Cruz, "5G Waveforms for
IoT Applications," in IEEE Communications Surveys & Tutorials, vol. 21, no. 3, pp. 2554-2567,
thirdquarter 2019, doi: 10.1109/COMST.2019.2910817.
[30] Proakis, John G, and Masoud Salehi. Digital Communications. McGraw-Hill, 2008. Print.
[31] Vu Tran Hoang Son, Dang Le Khoa, “BER Analysis for Downlink MIMO-NOMA Systems over
Rayleigh Fading Channels” in International Journal of Computer Networks & Communications,
vol.13, no.6, pp.93-108, November 2021, doi: 10.5121/ijcnc.2021.13606

89
AUTHORS
Kiran V. Shanbhag received his BE degree in Electronics and Communication
Engineering from Anjuman Institute of Technology and Management, Bhatkal,
Karnataka, India in 2005 and received his MTech degree in Communication
Engineering from National Institute of Technology Karnataka, Surathkal, India in 2010.
He is currently pursuing PhD in Department of Electronics and Communication
Engineering at St Joseph Engineering College, Mangaluru, India. His research interests
include digital signal processing for communication, multicarrier modulation schemes and MIMO radar.
He has 4 journal papers and 3 conference papers to his credit. Currently he is assistant professor at
department of electronics and communication engineering, Anjuman Institute of Technology and
Management, Bhatkal, Karnataka, India.
Savitha H. M. received her BE degree in Electronics and Communication Engineering
from P.E.S. College of Engineering, Mandya, Karnataka, India in 1987 and received her
MTech degree in Digital Electronics and Communication from NMAMIT, Nitte,
Karnataka, India in 2001. She got her PhD in Communication Engineering from
National Institute of Technology Karnataka, Surathkal, India in 2014. She was
associated with St Joseph Engineering College, Mangaluru as Professor and has a
teaching experience of 22 years. Her research interests include error control coding, OFDM, wireless
communication and signal processing. She has 7 international journals and 13 international/national
conferences to her credit.

Constellation Shared Multiple Access - A Noma Scheme for Increased User Capacity in 5G MMTC

More Related Content

Similar to Constellation Shared Multiple Access - A Noma Scheme for Increased User Capacity in 5G MMTC (20)

More from IJCNCJournal (20)

Recently uploaded (20)

Constellation Shared Multiple Access - A Noma Scheme for Increased User Capacity in 5G MMTC