Lte tutorial

1
3G Long-Term Evolution (LTE)
Farooq Khan
Samsung Telecom R&D Center
Richardson, Texas
September 25, 2006

2
LTE
• LTE is Evolution path for GSM/UMTS systems
• GSM is the dominant wireless cellular standard with over 2 Billion subscribers
Worldwide
– 82% of the global mobile market
UMTSUMTS
HSPAHSPA LTELTEGSMGSM
2005/2007 2010 and beyond
2Mb/s 100Mb/s14Mb/sKb/s

3
Tutorial Outline (1/2)
• Overview of LTE System Requirements
• Uplink multiple Access
– Orthogonal vs. Non-orthogonal multiple access
– Single Carrier FDMA (SC-FDMA) – DFT-Spread OFDM
– Spectral shaping for DFT-Spread OFDM
– Uplink frame structure and parameters
• Downlink Multiple Access
– OFDM overview
– OFDM vs. WCDMA
– Frequency diversity and frequency-selective multi-user scheduling
– Downlink frame structure and parameters
• Link Adaptation and Hybrid ARQ
– Frequency-domain link adaptation in OFDMA/SC-FDMA
– Channel frequency-selectivity based link adaptation
– Synchronous vs. Asynchronous and Non-adaptive vs. Adaptive Hybrid ARQ
– Chase combining vs. Incremental Redundancy (IR)
• Inter-cell interference mitigation techniques
– Fractional frequency reuse
– Fractional loading
– Interference suppression/cancellation

4
Tutorial Outline (2/2)
• Enhanced Multimedia Broadcast and Multicast Service
– Single-Frequency Network (SFN) operation
– Layered QoS (Quality of Service)
– Broadcast/Unicast Multiplexing
– Broadcast/Unicast Superposition and Interference cancellation
– E-MBMS frame structure
• MIMO Techniques
– MIMO capacity and Rank Adaptation
– Multi code word (MCW) MIMO
– Switching between Single-user and Multi-user MIMO
– Transmit diversity and beamforming
– MIMO for broadcast
• Flexible bandwidth support
– UE capabilities
– Synchronization and cell search
• LTE Architecture
– Network architecture
– Protocols architecture
• LTE Standard Development Schedule

5
LTE System Requirements
• Downlink peak data rate of 100Mb/s in 20 MHz (5 bps/Hz) and an
uplink peak data rate of 50Mb/s (2.5 bps/Hz).
• Spectrum efficiency, 2-4X Release 6 HSPA
• Scalable bandwidth support, [1.25] [1.6] 2.5, 5, 10, 15 and 20MHz
• Possibility for a radio-access network latency below 10 ms.
• Increase "cell edge bitrate" whilst maintaining same site locations as
deployed today.
• System should be optimized for low mobile speed but also support
high mobile speed
• Support for inter-working with existing 3G systems and non-3GPP
specified systems
• Efficient support of the various types of services, especially from the
PS domain (e.g. VoIP)
• Operation in paired and unpaired spectrum should not be precluded
Source: 3GPP TR 25.913

7
Uplink Multiple Access
• Design of an efficient multiple access and multiplexing
scheme more challenging on the uplink due to many-to-
one nature of the uplink transmissions.
• Two major approaches
– Non-orthogonal Access
• E.g. WCDMA
– Orthogonal Access
• E.g. OFDMA, TDMA, SC-FDMA etc.
• Peak-to-average power ratio (PAPR)
– Important aspect for the uplink due to limited UE transmit power

8
Conventional CDMA
F. Khan, “A Time-Orthogonal CDMA High Speed Uplink Data Transmission scheme for 3G and Beyond,” IEEE
Communication Magazine, February 2005.
F. Khan, “Performance of Orthogonal Uplink Multiple Access for Beyond 3G/4G Systems,” VTC 2006-Fall conference.
• It is advantageous to schedule good users in a TDM fashion and weak users in a
CDMA fashion in order to maximize the system capacity
• TDMA approach suffers from link budget limitation due to limited UE transmit power
i.e. a single user transmitting in a TDM fashion over a large bandwidth such as 5,
10, 20MHz may not be able to efficiently use the whole bandwidth due to its
transmit power limitation.
]//[
)1()1(
1log
0
2 Hzsb
NPKPf
P
KCCDMA 





+−++
+⋅=
α
Where f is ratio between other-cell and own-cell
signal and α is fraction of the own-user
signal considered as interference. For
the special case where f=0 and α=0, the
above equation simplifies to:
For large K, i.e.
]//[
)1(
1log
0
2 Hzsb
NPK
P
KCCDMA 





+−
+⋅=
]//[44.1)(log2 HzsbeCCDMA =≈
∞→K

9
Orthogonal vs. Non-orthogonal Access (1/3)
Where βι is the fraction of bandwidth allocated to
user i. For the case where the bandwidth
is equally divided among the K users
transmitting simultaneously, the above
formula can be simplified as below:
]//[1log
0
2
1
Hzsb
NfP
P
C
i
K
i
iOFDMA 





+
+⋅= ∑= β
β
]//[1log
0
2 Hzsb
NfKP
KP
COFDMA 





+
+=
We note that gains of orthogonal access over non-orthogonal access for larger SNR user case increase as the number of users increase.
the performance of high SNR user is dominated by intra-cell (or inter-user) interference and an orthogonal access benefits by eliminating the
intra-cell interference.
However, the performance of a weak user is dominated by the inter-cell interference and the background noise and eliminating intra-cell
interference by using orthogonal access only provides small advantage.
It should be noted that performance of a non-orthogonal scheme for larger SNR case (SNR=10.0dB) degrades as the number of users
increase. This is explained by the fact that increasing number of users also result in increased intra-cell interference.

10

11
Other-cell to own-cell
signal level ratio (f)
(COFDMA
- CCDMA
)/ CCDMA
[%]
Single user
SNR=0.0dB
Single user
SNR=10.0dB
0.0 151.59 342.65
0.2 83.19 103.02
0.5 51.97 58.41
0.8 38.20 41.47
1.0 32.53 34.85
• Note that this performance represents the potential gains that an orthogonal
scheme provides by eliminating multiple-access interference (MAI) on the
uplink.
• Orthogonal MA scheme such as OFDMA and SC-FDMA provides potential
additional gains in frequency selective channels over non-orthogonal WCDMA
using RAKE.

12
DFT-Spread OFDM (SC-FDMA)
IFFT
N
FFT
M
Data Data
Pilots
FFT
N
IFFT
M
Data
FDE
Data
Pilots
Equalized
Data
TRANSMITTER RECEIVER
Modulation symbols are FFT-pre-coded before mapping to the input of FFT
At the receiver, frequency-domain equalization is performed after the FFT
operation.
DFT-Spread OFDM is a relatively low PAPR waveform.
Approximately 3dB lower PAPR than OFDMA
DFT-Spread OFDM suffers from approximately 1dB link performance loss for
higher order modulations (e.g. 16-QAM) relative to OFDMA in a frequency-
selective channel.

13
Distributed FDMA
IFFT
N=8
FFT
M=4
a1
a2
a3
a4
b1
0
b2
0
b3
0
b4
0
a1
a2
a3
a4
a1
a2
a3
a4
a1, a2, a3, a4
M=4
a1, a2, a3, a4, a1, a2, a3, a4
S/P
P/S
Input signal
repeated N/M (=2)
times
time-domain signal
• Also referred to as IFDMA (Interleaved Frequency Division Multiple Access)*:
– Time-domain implementation
• In the frequency-domain implementation, the FFT coded data is mapped to equally
spaced subcarriers:
– FFT and IFFT operations “cancel” each other
– N/M repetitions at the output of IFFT (IFDMA principle)
• Distributed FDMA provides frequency-diversity. However, the performance is poor
with realistic channel estimation.
• Limited flexibility from resource allocation perspective.
*U. Sorger, I. De Broeck and M. Schnell,”Interleaved FDMA – A New Spread-Spectrum
Multiple-Access Scheme,” Proc. of ICC'98, Atlanta, Georgia, June 1998, pp. 1013-1017.

14
Localized FDMA
IFFT
N=8
FFT
M=4
a1
a2
a3
a4
b1
b2
b3
b4
0
0
0
0
a1
?
a2
?
a3
?
a4
?
a1, a2, a3, a4
M=4
a1, ?, a2, ?, a3, ?, a4 , ?
S/P
P/S
• The FFT coded modulation symbols are mapped to contiguous subcarriers
– FFT and IFFT operations do not “completely” cancel each other
– The input data is embedded (every (N/M)th symbol) in the stream at the output of
IFFT
• Greater flexibility in resource allocation
• Allows using channel sensitive scheduling
• Localized FDMA is preferred for LTE by most companies.

15
DFT-Spread OFDM PAPR
DFT-Spread OFDM provides approximately 3dB lower PAPR at 0.1%
point relative to an OFDM system.
Potential for larger coverage

16
Spectral Shaping (1/3)
• The PAPR of a DFT-spread OFDM system can be further reduced by
spectral shaping of the FFT-pre-coded data before mapping to the IFFT
input
• However, spectral shaping results in reduced spectral efficiency because of
the spectral shaping filter transition band.
– Examples of spectrum shaping filters are Raised Cosine Nyquist filter, Gaussian,
Hamming and Hann filters etc.
• Spectral shaping can further increase the system coverage because the cell
edge users are power limited (not bandwidth limited).
IFFT
N
FFT
M
Data
FFT
N
IFFT
M
Data
FDE
Data
Pilots
Equalized
Data
TRANSMITTER
RECEIVER
Spectral
Shaping

17
• The roll-off factor, α for a raised cosine Nyquist filter determines the excess
bandwidth.
• With M samples input to the spectrum shaping filter, the number of samples
at the output of the filter is M(1+ a). For example, if M=64 and a=0.25, the
number of samples at the output of the filter would be 64(1+0.25)=80.
• Therefore, 80 subcarriers would be required to map these samples at the
input of the IFFT.
– This represents 25% excess bandwidth.
Mapping to
subcarriers
FFT
M
Data M samples
M(1+α)
samples IFFT
N
0
0
0
0
N-M(1+α)
zeros
RF

18
Bandwidth
Efficiency
[bits/subcarrier]
Modulation RRC Roll-off 0.1%PAPR
[dB]
2 QPSK 0.0 5.7
1.6 QPSK 0.25 4.8
1.285 QPSK 5/9=0.5566 3.5
1 QPSK 1.0 1.8
1 π/2-BPSK 0.0 4.5
0.8 π/2-BPSK 0.25 2.8
0.64 π/2-BPSK 5/9=0.5566 1.0
0.5 π/2-BPSK 1.0 1.2
π/2-BPSK is a low-PAPR modulation scheme
However, when spectral shaping is employed in DFT-Spread OFDM, QPSK always
outperform π/2-BPSK in terms of PAPR for a given spectral efficiency.

19
Uplink Structure
• Subcarrier spacing, Long Block (LB)=15KHz, short Block (SB)=30KHz
• SB is used for pilot or reference signal transmission
• Uplink TTI (Transmission time Interval) consists of two subframes (1.0ms)
LB
(15KHz)
C
P
SB
30
KHz
C
P
LB
(15KHz)
C
P
LB
(15KHz)
C
P
LB
(15KHz)
C
P
LB
(15KHz)
C
P
SB
30
KHz
C
P
LB
(15KHz)
C
P
66.6us 33.3us
Subframe=0.5ms
Bandwidth
Frame=10ms
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Subframe=0.5ms TTI=1ms

20
Uplink Parameters
*1: {(x1/y1) × n1, (x2/y2) × n2} means (x1/y1) for n1 reference signal or data blocks and (x2/y2) for n2
reference signal or data blocks
*2: FFT size = samples
Source 3GPP TR 25.814
Spectrum
Allocation
(MHz)
Sub-frame
duration
(ms)
Long block size
(µs/#of occupied
subcarriers
/samples*2
)
Short block size
(µs/#of occupied
subcarriers
/samples)
CP duration
(µs/samples *1
)
20 0.5 66.67/1200/2048 33.33/600/1024 (4.13/127) × 7,
(4.39/135) × 1*
15 0.5 66.67/900/1536 33.33/450/768 (4.12/95) × 7,
(4.47/103) × 1*
10 0.5 66.67/600/1024 33.33/300/512 (4.1/63) × 7,
(4.62/71) × 1*
5 0.5 66.67/300/512 33.33/150/256 (4.04/31) × 7,
(5.08/39) × 1*
2.5 0.5 66.67/150/256 33.33/75/128 (3.91/15) × 7,
(5.99/23) × 1*
1.25 0.5 66.67/75/128 33.33/38/64 (3.65/7) × 7,
(7.81/15) × 1*1

22
OFDM (1/2)
• Orthogonal Frequency Division Multiple Access (OFDMA) promise higher
spectral efficiency by providing orthogonality between overlapping subcarriers.
• A guard interval or cyclic prefix is added to combat the multipath delay spread.
…
Sub-carriers
FFT
Time
Symbols
5 MHz Bandwidth
Guard Intervals
…
Frequency

23
OFDM (2/2)
• OFDM requires a single IFFT operation at the transmitter and a single FFT operation at
the receiver.
• In contrast, DFT-Spread OFDM requires two FFT/IFFT operations at the transmitter
and two FFT/IFFT operations at the receiver.
IFFT
TRANSMITTER
RECEIVER
P/SS/P
Mod.
Symbols
Add
CP
RF
S/P FFT
Rem
ove
CP
Receive P/S
Mod.
Symbols
QAM
Mod.
Coded bits
QAM
Dem
od.
to decoder

24
OFDM vs. WCDMA (1/2)
The peak data rate achievable with WCDMA will be limited to
1b/s/Hz when the unicast signal is received with large
number of multi-path components.
where ρ((≡P/N0) is the SINR when all the power is received on
a single-path and there is no interference from the
other cells. f represents the ratio between other-cell
and own-cell signal.
In OFDM, there is no multi-path interference due to use of a
cyclic prefix and 1-tap equalization of OFDM
subcarriers. Therefore, the only sources of SINR
degradation in an OFDM system are the other-cell
interference and the background noise. The SINR in
an OFDM system can therefore be expressed as:
Where Ts is the OFDM symbol duration and GI is the cyclic
prefix duration.
( )
]//[
11
1log2 Hzsb
f
CWCDMA 





++⋅
+=
ρ
ρ
]//[
1
1log2 Hzsb
fGIT
T
C
s
s
OFDMA 





+⋅
+⋅





+
=
ρ
ρ
f K=2 K=4 K>>1
0.0 122% 180% 237%
1.0 21% 36% 51%
2.0 9% 18% 27%

25
OFDM vs. WCDMA (2/2)
• OFDM provides greater capacity
advantage for the cases when
the other-cell interference (f) is
relatively small.
– users closer to the base station
with good channel quality
benefits more from OFDM
(performance dominated by the
multi-path interference).
• For weak users at the cell edge
(f relatively larger), the
performance is dominated by
interference from neighbouring
cell rather than the own-cell-
interference. Therefore, OFDM
provides little advantage for the
weak users in the cell.
f K=2 K=4 K>>1
0.0 23% 40% 55%
1.0 10% 19% 28%
2.0 4% 11% 17%

26
OFDM vs. WCDMA w/ advanced receiver
• OFDM outperforms WCDMA with both RAKE and with MMSE receiver.
• However, the performance difference between OFDM and WCDMA with MMSE
is relatively smaller.
– Performance of WCDMA improves significantly using an MMSE receiver compared to
a RAKE receiver.
Performance of WCDMA vs. OFDM Unicast:
Mixed Ped A and B at 50:50, 3km/h, Proportional Fair scheduler
0.0
1.0
2.0
3.0
4.0
5.0
6.0
7.0
0 5 10 15 20 25 30
Number of users per sector
SectorThroughput[Mbps]
OFDM
WCDMA(RAKE)
WCDMA(MMSE)

27
Frequency-diversity and scheduling
• In case of frequency-selective multi-user scheduling, a contiguous set of subcarriers
potentially experiencing an upfade is allocated for transmission to a user.
– generally beneficial for low mobility users for which the channel quality can be tracked.
• in case of frequency-diversity transmission, the allocated subcarriers are preferably
uniformly distributed over the whole spectrum
– Generally used for high mobility users for which channel quality can not be tracked
– Also beneficial for control signalling and delay-sensitive services
Distributed subcarrier Allocation
for Frequency diversity
f
1
f
2
f
3
f
4
f
5
f
6
f
7
f
8
f
9
f
1
0
f
1
1
f
1
2
f
1
3
f
1
4
f
1
5
f
1
6
f
Contiguous subcarrier Allocation for
Frequency-Selective Multi-user scheduling
f
1
f
2
f
3
f
4
f
5
f
6
f
7
f
8
f
9
f
1
0
f
1
1
f
1
2
f
1
3
f
1
4
f
1
5
f
1
6
f

28
Frequency-selective multi-user scheduling
• The overall signal quality can be improved if user 1 is scheduled at the edge
resource blocks where user1’s signal quality is better and user2 in the middle
resource blocks where user2’s signal quality is better.
• Frequency-selective multi-user scheduling enabled by OFDM transmission can
provide additional performance benefit over WCDMA system.
FrequencyOFDM Subcarriers
Flat
Fading
User1Signal
Power
User2
Frequency
Selective
Fading
RB#
1
RB#
2
RB#
3
RB#
4
RB#
5
RB#
6
RB#
7
RB#
8
RB#
9
RB#
10
RB#
11
RB#
12

29
OFDM w/ Frequency-selective scheduling
• Frequency-selective multi-user scheduling provides additional throughput gain
on top of multi-user diversity using time-domain scheduling.
• The channel quality feedback overhead is larger for a system employing
frequency-selective multi-user scheduling.
Frequency Scheduling Gain in OFDMA with Contiguous Structure:
Channel Model: Ped B
3.0
4.0
5.0
6.0
7.0
8.0
0 5 10 15 20 25 30
Number of users per sector
SectorThroughput[Mbps]
# of sub-bands: 1
# of sub-bands: 8

30
Downlink Frame Structure
The downlink TTI consists of two 0.5ms subframes (1ms)
A 10ms frame consists of 20 subframes (10 TTIs)
A subframe contains 7(6) OFDM symbols with short (long)
cyclic prefix (CP).
Pilot or reference signals are transmitted in the 2 OFDM
symbols every 6th
subcarrier per antenna.
Pilot for multiple transmit antennas are frequency-
multiplexed (FDM).
Subframe=0.5ms
P1
D
D
P2
D
D
P1
D
D
P2
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
P2
D
D
P1
D
D
P2
D
D
P1
D
D
D
D
D
D
D
D
D
D
D
D
D
D
Frequency
D
D
D
D
D
D
D
D
D
D
D
D
P1 Pilot for ANT1 P2 Pilot for ANT2
D Data or Control Symbols
Frame=10ms
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Subframe=0.5ms TTI=1ms

31
Downlink Parameters
Transmission BW 1.25 MHz 2.5 MHz 5 MHz 10 MHz 15 MHz 20 MHz
Sub-frame duration 0.5 ms
Sub-carrier spacing 15 kHz
Sampling frequency 1.92 MHz
(1/2 × 3.84 MHz)
3.84 MHz 7.68 MHz
(2 × 3.84 MHz)
15.36 MHz
(4 × 3.84 MHz)
23.04 MHz
(6 × 3.84 MHz)
30.72 MHz
(8 × 3.84 MHz)
FFT size 128 256 512 1024 1536 2048
Number of occupied
sub-carriers†, ††
76 151 301 601 901 1201
Number of
OFDM symbols
per sub frame
(Short/Long CP)
7/6
Short (4.69/9) × 6,
(5.21/10) × 1*
(4.69/18) × 6,
(5.21/20) × 1
(4.69/36) × 6,
(5.21/40) × 1
(4.69/72) × 6,
(5.21/80) × 1
(4.69/108) × 6,
(5.21/120) × 1
(4.69/144) × 6,
(5.21/160) ×1
CP length
(μs/samples)
Long (16.67/32) (16.67/64) (16.67/128) (16.67/256) (16.67/384) (16.67/512)
†Includes DC sub-carrier which contains no data
†† This is the assumption for the baseline proposal. Somewhat more carriers may be possible to occupy
in case of the wider bandwidth
*: {(x1/y1) × n1, (x2/y2) × n2} means (x1/y1) for n1 OFDM symbols and (x2/y2) for n2 OFDM symbols

32
Link Adaptation and Hybrid ARQ

33
Link Adaptation and Hybrid ARQ
• Link adaptation is used to adapt the modulation and coding to the
instantaneous channel conditions reported by the UE.
• Hybrid ARQ operates on top of link adaptation to recover from
errors in MCS selection.
– MCS selection errors occur due to CQI (channel quality indication)
errors
– CQI errors are a result of channel quality measurement inaccuracy,
feedback delay and interference fluctuations etc.

34
Frequency-domain link adaptation (1/2)
• In frequency-domain link adaptation (LA), different modulations can be used for coded bits
mapped to different OFDM subcarriers experiencing different channel quality.
• In DFT-Spread OFDM, frequency-domain link adaptation does not apply because a given
modulation symbol is spread over the whole transmitted bandwidth.
Flat
Fading
Signal
Power Frequency Selective
Fading
16-
QAM
16-
QAM
16-
QAM
16-
QAM
QP
SK
QP
SK
QP
SK
16-
QAM
16-
QAM
16-
QAM
16-
QAM
16-
QAM

35
Frequency-domain link adaptation (2/2)
• LTE system does not employ frequency-domain link adaptation:
– A common modulation is selected for all the coded symbols transmitted.
Transport block (L2 PDU)
CRC attachment
Channel coding
HARQ functionality
including adaptive
coding rate
Physical channel
segmentation
(resource block mapping)
Adaptive modulation
(common modulation is selected)
To assigned resource blocks
Number of assigned
resource blocks

36
Channel selectivity based LA (1/2)
0 1 2 3 4 5 6 7 8
10
-4
10
-3
10
-2
10
-1
10
0
EbNo(dB)
FER
QAM16, 1/3, Flat
QPSK, 2/3, Flat
QAM16, 1/3, TU-6-path
QPSK, 2/3, TU-6-path
• Modulation switching to next higher level of modulation happens at a relatively low
coding rate when the OFDM transmission experience frequency-selective fading.
• On the other hand, the switching to next higher level of modulation happens at a
relatively higher coding rate when the OFDM transmission experiences a “flat-fading”.

37
Channel selectivity based LA (2/2)
• In a flat-fading channel QPSK can be used for as high coding rate as 5/6 before
switching to 16-QAM.
• In a frequency-selective channel such a GSM Typical Urban (TU) channel, switching
from QPSK to 16-QAM happens at around ½ coding rate.
• Possible different modulation switching points are required for frequency-selective multi-
user scheduling mode and frequency diversity mode.
0 1 2 3 4 5 6 7 8
10
-4
10
-3
10
-2
10
-1
10
0
EbNo(dB)
FER
QAM16, 3/8, Flat
QPSK, 3/4, Flat
0 2 4 6 8 10 12
10
-4
10
-3
10
-2
10
-1
10
0
EbNo(dB)
FER
QAM16, 5/12, Flat
QPSK, 5/6, Flat

38
Hybrid ARQ
SP1
Packet
P
Channel
Coding
Subpacket
Generator
SP2 SP3 SP4
RECEIVER
NACK
TRANSMITTER
SP1
NACK
SP2
ACK
SP3
SP1, SP2
Combine and
Decode
SP1 Decode
SP1, SP2, SP3
Combine and
Decode

39
Sync. Vs. Async. Hybrid ARQ
• Asynchronous Hybrid ARQ allows to schedule retransmissions at favorable
channel and resource conditions at the expense of additional control
overhead.
• LTE system employs an asynchronous Hybrid ARQ on the downlink and a
synchronous scheme in the uplink.
1
SP1
2 3 4
5
SP2
6 7 8
9
SP2
10 11 12
13
SP4
14 15 16
CT
RL
NACK
NACK
NACK
ACK
1
SP1
2 3 4 5 6
7
SP2
8 9 10
11
SP3
12
13
14 15
16
SP4
CT
RL
CT
RL
CT
RL
CT
RL
NACK
NACK
NACK
Synchronous
Asynchronous

40
Chase Vs. IR Hybrid ARQ
• In chase combining, a copy of the original transmission is retransmitted on a
NACK response:
– Retransmissions provide SNR and diversity gain
• In Incremental Redundancy (IR), the retransmissions potentially contain
redundant coded bits
– Retransmissions also provide additional coding gain
• IR leads to better performance at the expense of additional complexity due
to buffering
1
SP1
2 3 4 5 6
7
SP2
8 9 10
11
SP3
12 13 14 15
16
SP4
NACK
NACK
NACK
Incremental Redundancy
1
SP1
2 3 4 5 6
7
SP1
8 9 10
11
SP1
12 13 14 15
16
SP1
Chase Combining
NACK
NACK
NACK

41
Adaptive Vs. Non-Adaptive Hybrid ARQ
• In non-adaptive Hybrid ARQ, the modulation, coding and resources
allocated for retransmissions is the same as the original
transmission.
• An adaptive Hybrid ARQ scheme allows to adapt the modulation,
coding and resources allocated for retransmissions based on
channel conditions and/or other criteria.
1
SP1
2 3 4 5 6
7
SP2
8 9 10
11
SP3
12 13 14 15
16
SP4
NACK
NACK
NACK
Adaptive
1
SP1
2 3 4 5 6
7
SP1
8 9 10
11
SP1
12 13 14 15
16
SP1
Non-adaptive
NACK
NACK
NACK
16-QAM 16-QAM 16-QAM 16-QAM
16-QAM 16-QAM QPSK QPSK

42
Low overhead Async. Hybrid ARQ
• Control information is only transmitted if one or more of the subpacket
timing, modulation, coding, transmission duration or resource information
needs to be changed.
1
SP1
2 3 4
5
X
6 7 8
9
Y
10 11
12
SP2
13 14 15
16
SP3
CT
RL
CT
RL
SP1 SP1, X SP1, X, Y SP1, SP2
SP1,
SP2, SP3
Discard
X and Y
Success
TX
RX
NACK
NACK
NACK
NACK
ACK

43
Inter-Cell Interference Mitigation
Techniques

44
Fractional Frequency Reuse
In the inner cell region, a universal frequency
reuse (reuse of 1) is employed.
For the cell-edge users, a reuse > 1 is used.
A higher power spectral density (PSD) is used
on the subcarriers allocated to the cell-
edge users.
- cell-edge users are generally power
limited
A lower power spectral density is used on the
subcarriers allocated to the users closer
to the base station experiencing good
channel conditions.
- good users are generally bandwidth-
limited (not power limited).
Fractional frequency reuse may have negative
impact on frequency-selective multi-user
scheduling.
f1
f1
f1
f2
f4
f3
Low PSD
High PSD
RB
1
RB
2
RB
3
RB
4
RB
6
RB
7
RB
8
RB
9
RB
11
RB
12
RB
5
RB
10
f1 f2 f3 f4
Power
Total OFDM Bandwidth

45
Cell-edge Performance
We assume a pathloss exponent of α=3.76 and an
interference-limited scenario therefore ignoring the
effect of background thermal noise.
A reuse of 3 can provide approximately 2X improvement
(1.07 vs. 0.51 b/s/Hz) in cell-edge performance
relative to the case of universal frequency reuse.
These gains apply to channels targeted for the cell-edge
such as broadcast and paging channels.
For data traffic, generally the improvement in cell-edge
performance via inter-cell interference
coordination results in decrease in overall system
throughput.
57
1
6
4
3
2
13
14
15
12
11
10
9
8
19
18
17
16
( ) ( ) ( )[ ] ( ) ( )
dB
RRR
R
reuse 7.3424.0
136232
1
136232
1 −==
+×+×+
=
+×+×+×
= −−−−−
−
− ααααα
α
ρ
( ) ( )[ ] ( ) ( )
dB
RR
R
reuse 22.9368.8
3
13623
1
3
13623
3 ==
+×+×
=
+×+×
= −−−−
−
− αααα
α
ρ
( ) ( )
( ) ( ) ]//[07.1368.81log
3
1
1log
3
1
]//[51.00.4241log1log
2323
2121
HzsbC
HzsbC
reuseresue
reusereuse
=+×





=+×





=
=+=+=
−−
−−
ρ
ρ

46
Fractional Loading (1/2)
In the fractional loading approach, each cell operates at a duty cycle smaller than 100% on a given time-frequency
resource.
The average capacity in a fractional loading approach can be approximated as:
Where ci is the capacity in a time-frequency resource with i number of transmissions among the neighbouring cells.
0.5ms 0.5ms 0.5ms
0 1 2 1 0 0
TotalBandwidth
Cell-A Cell-B Cell-C
RB#1
2
3
4
5
6
7
8
9
10
11
12
( ) ]//[1log
]//[
2
33221100
Hzsbc
HzsbcpcpcpcpC
ii
average
ρ+=
+++=

47
Fractional Loading (2/2)
Fractional loading approach can provide around 55% (0.87
vs. 0.56 b/s/Hz) improvement in cell-edge
performance relative to reuse of 1 (universal
reuse).
In addition, under an ideal frequency reuse of 3,
performance can be further improved by
approximately 20% (1.07 vs. 0.87 b/s/Hz) relative
to fractional loading. However, it should be noted
that a frequency reuse approach might be
challenging to implement in practice due to
required frequency planning etc.
ρi indicates the SINR experienced with i number of
transmissions among the neighbouring cells and
are approximated as:
( ) ( )[ ] ( ) ( )
( ) ( ) ( )[ ] ( ) ( )
( ) ( ) ( )[ ] ( ) ( )
dB
RR
R
R
dB
RR
R
R
dB
RR
R
dB
26.347.0
3
13623
2
1
3
13623
2
49.089.0
3
13623
1
1
3
13623
22.9368.8
3
13623
1
3
13623
0.0
3
2
1
0
−==
+×+×
+
=
+×+×
+×
=
−==
+×+×
+
=
+×+×
+
=
==
+×+×
=
+×+×
=
−∞==
−−−−
−
−
−−−−
−
−
−−−−
−
αααα
α
α
αααα
α
α
αααα
α
ρ
ρ
ρ
ρ

48
Interference Suppression/Cancellation
With multiple receive antennas, the mobile station can
potentially suppress neighbouring sector/cell
interferers.
With 2-Rx and non-MIMO (rank1) transmission, a
single dominant interferer can be suppressed by
using interference rejection combining (IRC).
With rank2 MIMO transmission (2 streams), the
dominant interferer for a given MIMO stream is
the other MIMO stream(s) and hence the
receiver degree of freedom is used to suppress
the other streams.
Therefore, in a MIMO transmission, IRC is not likely to
provide any gains.
3-sector-BS
1
MS
2
3
In inter-cell interference cancellation (IIC), the interferer signal is decoded and then cancelled.
IIC can provide significant gains (up to 3dB SINR gain) in cell-edge performance improvement.
However, IIC has some practical limitation due to the need of decoding control information
along with data transmissions from the interfering cells.

49
Multimedia Broadcast and Multicast
Service

50
Single-Frequency Network (SFN)
A single-frequency network (SFN) operation can be realized
for broadcast traffic transmitted using OFDM from
multiple synchronized cells (with timing errors within
the cyclic prefix length).
When the same information content is transmitted using
identical modulation, coding and time-frequency
resources, the signal is “RF” combined and there is no
interference apart from the background noise.
The pilot for SFN operation is also transmitted using identical
pilot sequence and time-frequency resources.
In the presence of SFN operation, the broadcast SINR can be
very high particularly for smaller cells deployments.
At 90% coverage point, SFN provides SINR of 11, 21 and
28dB for cell site-to-site distance of 2, 1 and 0.5Km
respectively.
3
1
2
7
17
18 10
19
8
9
4
5
6
11
14
16
13
12
15

51
SFN vs. WCDMA for Broadcast
We observe that the achievable capacity for broadcast in a
WCDMA system is limited to 1b/s/Hz when the
broadcast signal is received from large number of base
stations (K>>1).
In case of OFDM, the signals received from multiple
synchronized base stations are orthogonal as long as
the relative delays of the received signals are within the
OFDM symbol cyclic prefix length.
Where Ts is the OFDM symbol duration and GI is the cyclic prefix
duration.
The Single Frequency Network (SFN) gains are larger in smaller
cells where strong signals are received from multiple
cells by the cell edge users.
]//[
)1(
1log
0
2 Hzsb
NPK
KP
CWCDMA 





+−
+=
]//[1log
0
2 Hzsb
N
KP
GIT
T
C
s
s
OFDMA 





+⋅





+
=
Broadcast
Geometry, KP/N0
[dB]
Capacity gain of OFDM over WCDMA
[(COFDM-CWCDMA)/CWCDMA*100]
0.0dB 49%
7.0dB 152%
10.0dB 215%

52
Layered QoS via Hierarchical Modulation
The layered QoS concept based on hierarchical modulation is already in use in DVB (Digital Video
Broadcast) and MediaFLO systems.
A high priority stream is embedded within a low priority stream. The users with good signal reception
quality receive both streams while the users with poor channel conditions only receive the
high priority stream.
The good channel quality users can have an enhanced video reception quality due to higher data
rates reception of the same content.
Weak users decode
signal as QPSK
'00'
Good users decode
signal as 16-QAM
'0010'
10
01 00
11

53
Broadcast/Unicast Multiplexing
Frequency multiplexing of broadcast/unicast provides the flexibility to adaptively allocate power
between broadcast and unicast traffic.
In an interference-limited scenario, unicast power can be reduced without affecting unicast
throughput while increasing the broadcast power which results in linear increase in SINR in
an SFN scenario.
Unicast
(alpha*P)
0.5ms
Unicast
(alpha*P)
Unicast
(alpha*P)
Broadcast
(P)
Unicast
(alpha*P)
Unicast
(alpha*P)
Unicast
(alpha*P)
Broadcast
(P)
Wasted Power
(1-alpha)*P
Wasted Power
(1-alpha)*P
Ptotal
Unicast
(0.75*
alpha*P)
0.5ms
Unicast
(0.75*
alpha*P)
Unicast
(0.75*
alpha*P)
Unicast
(0.75*
alpha*P)
Unicast
(0.75*
alpha*P)
Unicast
(0.75*
alpha*P)
Broadcast, (1-0.75*alpha)*P, 1/4th of the subcarriers
Unicast
(0.75*
alpha*P)
Unicast
(0.75*
alpha*P)
Frequency-Multiplexing
Time-Multiplexing
Ptotal
Punicast

54
Broadcast/Unicast Multiplexing
LTE supports FDM/TDM transmission of unicast and broadcast/multicast traffic within the same RF
carrier.
TDM multiplexing of different MBMS streams (broadcast channels) is supported to minimize the
broadcast reception time at the UE.
A UE listening to a particular channel only needs to turn on its receiver during the TTIs transmitting
that channel resulting in battery power savings.
Unicast Unicast
Unicast
Broadcast
CH 2
Unicast
Unicast Unicast
TTI=1ms
Broadcast
CH 1
Unicast
Broadcast
CH 4
Unicast
Broadcast
CH 3
Frequency

55
Broadcast/Unicast Superposition (1/4)
In the Unicast/MBMS superposition approach,
the same time-frequency resources are
used for simultaneous transmission of
Unicast and MBMS traffic.
The proposed approach superimposes the
broadcast signal over the unicast traffic
and cancels the broadcast signal before
unicast demodulation and decoding.
In a single frequency network (SFN) based
MBMS transmission, interference from
all the cells in a broadcast zone is
cancelled in a single-step by cancellation
of the composite received MBMS signal.
The composite received signal is reconstructed
for cancellation purpose by using the
composite channel estimates based on
MBMS reference signal.
The broadcast signal uses excess unicast power
available in interference limited situation
for SFN operation therefore converting
the unused power into useful capacity.
Unicast/MBMS superposition approach also
applies to MIMO cases.
Ucast
TDM of MBMS and Unicast
Ucast Ucast Bcast Ucast Ucast Ucast Bcast
Ucast
Bcast
Ucast
Bcast
Ucast
Bcast
Ucast
Bcast
Ucast
Bcast
Ucast
Bcast
Ucast
Bcast
Ucast
Bcast
TIME
FREQUENCY
TIME
FREQUENCY
FDM of MBMS and Unicast
Ucast
Superposition of MBMS and Unicast
Ucast Ucast Ucast Ucast Ucast Ucast Ucast
TIME
FREQUENCY Bcast Bcast Bcast Bcast Bcast Bcast Bcast Bcast

56
57
1
6
4
3
2
57
1
6
4
3
2
+ =
57
1
6
4
3
2
57
1
6
4
3
2
Broadcast
(Same content in cells within
a broadcast zone)
Unicast
(different content
in different cells)
Broadcast/
Unicast
Superposition
57
1
6
4
3
2
57
1
6
4
3
2
- =
57
1
6
4
3
2
57
1
6
4
3
2
Decode and
Cancel
Broadcast
Unicast Signal
Broadcast/Unicast
Superimposed Received
Signal
Superposition at BS
Interference Cancellation at MS

57
• The MBMS signal is transmitted over the unicast resources without affecting unicast
performance while allowing for additional “free” capacity for MBMS.
• The unused unicast power is converted to useful capacity for broadcast.
• An additional 1.28 b/s/Hz spectral efficiency, can be provided for the cell when MBMS signal
is superimposed on unicast resources. This meets and exceeds the LTE MBMS spectral
efficiency requirement of 1.0 b/s/Hz without requiring an additional separate resources for
MBMS. This can be seen as MBMS traffic is carried for “free” meeting the LTE target.
Free MBMS
Capacity
Unused
Node-B
Power
Unicast
SFN
MBMS
Unicast Unicast
Cancelled
MBMS
signal
Total
Node-B
Power

58
Additional unicast traffic can be
transmitted in parallel with MBMS
on resources reserved for MBMS.
A significant unicast capacity
advantage can be obtained.
2X system throughput improvement for
unicast is achieved while at the
same time slightly improving the
MBMS throughout.
The gains from the Unicast/MBMS
superposition approach are
dependent upon the fraction of
the resources allocated for
unicast and MBMS.
If a given network only carriers either
unicast or MBMS but not both,
then there are no improvements
possible because unicast and
MBMS cannot be superimposed.
Full Unicast
Capacity
SFN
MBMS
SFN
MBMS
Unicast Unicast
Cancelled
MBMS
signal
Small reduction
in MBMS
capacity
Total
Node-B
Power
Unicast
15.2Mb/s 27.4Mb/s
MBMS
10MHz 10MHz
30.4Mb/s (1.52b/s/Hz) [Unicast]
30.8Mb/s (1.54b/s/Hz) [MBMS]
20MHz
Orthogonal (TDM/FDM) Multiplexing
Unicast/MBMS Superposition and Cancellation

60
Multiple Input Multiple Output (MIMO)
A MIMO system promises linear increase in capacity with the number of antennas.
Ant1
TRANSMITTER
Data Stream 1
Data Stream 2
Data Stream 3
Data Stream 4
Ant2
Ant3
Ant4
Ant1
Ant2
Ant3
Ant4
Spatial
Processing
RECEIVER
Data Stream 1
Data Stream 2
Data Stream 3
Data Stream 4

61
MIMO Capacity at Low SNR
The capacity of an MxM MIMO channel can be written as:
Where SNR is the received signal-to-noise ratio at each receive antenna.
At low SNR, an MxM system yields a power gain of M relative to a single-receive antenna case
because the receive antennas can coherently combine their received signals to get a power
boost.
]//[detlog *
2 HzsbHH
M
SNR
IEC MMIMO 











+=
]//[1log
1
2
2 Hzsb
M
SNR
EC
M
i
iMIMO ∑=












+= λ min21 λλλ ≥≥≥ 
( )[ ]
[ ][ ] [ ]
[ ]
[ ]HzsbeSNRM
HzsbehE
M
SNR
HzsbeHHTrE
M
SNR
eE
M
SNR
C
ji
ji
M
i
iMIMO
//log
//log
//log
log
2
2
,
2
'
2
*
1
2
2
××=






=
=
≈
∑
∑=
λ
( ) exx 22 log.1log ≈+
Source: “Fundamentals of Wireless Communication” by D. Tse and P. Viswanath

62
MIMO Capacity at High SNR
For the high SNR case where we can use the approximation, the MIMO capacity formula can be
expressed as:
the full M degree of freedom is attained.
It can be noted that maximum capacity is achieved when all the singular values are equal.
( )
( )[ ]∑
∑
∑
=
=
=
+





×=






+





=












=
M
i
i
M
i
i
M
i
iMIMO
HzsbE
M
SNR
M
Hzsb
M
SNR
E
Hzsb
M
SNR
EC
1
2
22
1
2
22
1
2
2
]//[loglog
]//[loglog
]//[log
λ
λ
λ ( ) ( )xx 22 log1log ≈+
( )[ ] −∞>2
2log iE λ
Source: “Fundamentals of Wireless Communication” by D. Tse and P. Viswanath

63
MIMO Rank Adaptation
The number of MIMO streams transmitted to a user are adapted based on user channel quality (CQI) and
channel matrix profile.
A Rank1 transmission is advantageous for weak users experiencing low SINR and also for high SINR users
with a rank deficient MIMO channel.
The users provide the rank feedback to the base station on a slow basis (on the order of 100ms)
Mobile Station
BS
Pilot (ANT1)
CQI + channel rank indication
Data
Calculate CQI and
channel rank
Pilot (ANT2)
Select transmission
rank and schedule user
on the selected
antenna (s)

64
MIMO Pre-coding
A precoding is used to create a set of virtual antennas from physical antennas.
With precoding, the full base station power (for all the antennas) can always be used irrespective of the
number of virtual antennas used for transmission.
Precoding with a reasonably large codebook size can also provide beamforming gains at the expense of
additional signaling overhead.
LTE employs unitary-precoding for single-user MIMO. For multi-user MIMO, both unitary and non-unitary
precoding approaches are under study.
P1
ANT1
P2
2
)( 21 SS +
ANT2
2
)( 21 SS −
ANT1
2
)( 21 SS +
ANT2
2
)( 21 jSjS −
1S
2S
1S
2S
VA1
VA1
VA2
VA2






−
=





−
=
jj
PP
11
2
1
,
11
11
2
1
21
inv(P1)
inv(P2)
2
)( 21 SS +
2
)( 21 SS −
2
)( 21 SS +
2
)( 21 jSjS −
1S
2S
1S
2S






−11
11
2
1





 −
j
j
1
1
2
1





 −
=





−
=
j
j
PinvPinv
1
1
2
1
)(,
11
11
2
1
)( 21

65
Single-user vs. Multi-user MIMO
In single-user MIMO, all the virtual antennas (or MIMO layers) on a given time-frequency resource are
allocated to the same user.
In multi-user MIMO, the virtual antennas (or MIMO layers) on a given time-frequency resource can be
shared by multiple users.
Single-user MIMO increases user peak data rate while multi-user MIMO increases system throughput.
LTE supports both single-user and multi-user MIMO.
FrequencySingle-User MIMO
S2
S1VA1
VA2 S2
S1
S2
S1
S2
S1
S2
S1
S2
S1
S2
S1
S2
S1
FrequencyMulti-User MIMO
S2
S1VA1
VA2 S2
S1
S2
S1
S2
S1
S2
S1
S2
S1
S2
S1
S2
S1
User-A User-B
User-A
User-B

66
SCW vs. MCW MIMO
LTE employs multi-codeword (MCW) single-user MIMO approach. MCW allows for per antenna
rate control (PARC) and successive interference cancellation (SIC).
MCW approach provides better performance relative to single codeword (SCW) at the expense of
additional signaling overhead.
Information
Turbo/
LDPC
coding
Modulation
(QPSK, 16-
QAM etc.)
Attach
CRC
ANT1
ANT2
Demux
DemuxInformation
Turbo/
LDPC
coding
Turbo/
LDPC
coding
Modulation
(QPSK, 16-
QAM etc.)
Modulation
(QPSK, 16-
QAM etc.)
Attach
CRC
Attach
CRC
ANT1
ANT2
S1
S2
S1
S1
SCW
MCW

67
SIC Receiver for Single-user MIMO
For MCW single-user MIMO, a successive interference cancellation (SIC) receiver can be used.
Decode
stream 1
Interference
cancellation
MMSE
Decode
stream 2
Interference
cancellation
MMSE
Decode
stream 3
Interference
cancellation
MMSE
Decode
stream 4
MRC
Received
Signal

68
Multi-user MIMO (1/2)
In multi-user MIMO, the virtual antennas (or
MIMO layers) on a given time-frequency
resource are shared by multiple users.
Multi-user MIMO exploits multi-user diversity in
the spatial domain.
Multi-user MIMO provides reasonable
performance with a simple LMMSE
receiver.
The channel quality indication (CQI) overhead
for MU-MIMO is lower relative to single-
user MIMO.
The multi-user MIMO does not increase the user
peak data rate.
Flat
Fading
User-1Signal
Power
User-2 Frequency
Selective Fading
RB1
(User-2)
RB1
(User-1)
ANT1
ANT2
ANT1
ANT2
RB2
(User-2)
RB2
(User-1)
RB3
(User-1)
RB3
(User-2)
RB4
(User-2)
RB4
(User-2)
RB5
(User-2)
RB5
(User-2)
RB6
(User-1)
RB6
(User-1)
RB7
(User-1)
RB7
(User-1)
RB8
(User-1)
RB8
(User-1)
User-1 User-2

69
Multi-user MIMO (2/2)
In multi-user MIMO, the user feedback the best virtual antenna CQI and the antenna selection
indication (1-bit for a 2x2 MIMO system).
RECEIVER
TRANSMITTER
Pilot (ANT1)
CQI + 1-bit ANT indication
Data
Select the greater of
CQI1 and CQI2
CQI=max(CQI1, CQI2)
Calculate CQI1 and
CQI2 assuming LMSSE
Pilot (ANT2)
Select MCS based on
CQI and schedule user
on the selected antenna

70
Single-user/Multi-user MIMO Switching
Both static and dynamic switching approaches between single-user and multi-user MIMO are proposed for
the LTE.
In the dynamic approach, the scheduler dynamically selects a mode of transmission that maximizes capacity
at a given time.
In general, multi-user MIMO outperforms single-user MIMO at higher loads. At light systems load, single-user
MIMO outperforms multi-user MIMO
1
B11
2 3 4
5
B21
6 7 8
9
A12
10 11 12
13
A21
14 15 16
A1 failure
B2 success
Stream-1
RX-A A1 success
A2 success
NACK
1
A11
2 3 4
5
B12
6 7 8
9
B22
10 11 12
13
A13
14 15 16
Stream-2
B1 failure
B1 success
B2 failure
A1 failureRX-B
NACK
NACK
ACK(1,1)
ACK
ACK(1,0)

71
Transmit Diversity Schemes
Both rank1 and rank2 transmit diversity schemes have been proposed for LTE.
Rank1 transmit diversity schemes include cyclic delay diversity (CDD), frequency-switched transmit
diversity (FSTD), time-switched transmit diversity (TSTD)
Rank2 based transmit diversity scheme is based on space frequency block code (SFBC)
Rank2 transmit diversity provides small link performance gain of over rank1 transmit diversity
schemes for the 2Tx-2Rx case depending upon the coding rate etc.
Rank1 transmit diversity schemes such as CDD and FSTD scale easily with the number of transmit
antennas. Rank1 schemes also has the advantage in inter-cell interference suppression and
joint operation with spatial multiplexing.
Transmit diversity scheme(s) for LTE is still an open issue.
S1
ANT1
S2
S1.eθ1
S2.eθ2
ANT2
S1
ANT1
NULL
NULL
S2
ANT2
S1
ANT1
-S2*
S2
S1*
ANT2
CDD FSTD SFBC
Frequency

72
MIMO for Broadcast
For broadcast traffic, only open-loop MIMO schemes
such as open-loop transmit diversity scheme
or an open-loop spatial multiplexing approach
can be considered.
Any form of additional transmit diversity is not
expected to bring any significant benefit
because SFN-based broadcast already enjoys
from frequency-diversity due to delayed
signals received from multiple cells.
Spatial multiplexing MIMO approach is attractive for
broadcast due to very high SINR experienced
and the fact that received signals from multiple
cells see increasing decorrelations in an SFN
environment.
Broadcast/
Multicast
Controller
Broadcast/
Multicast
Content
Server
Coding and
Modulation
DMUX
ANT1
ANT2
S1
S2
Coding and
Modulation
DMUX
ANT1
ANT2
S1
S2
Coding and
Modulation
DMUX
ANT1
ANT2
S1
S2
BS3
BS1
BS2
Same Broadcast
Multicast
Information
Receiver
ANT1
ANT2

74
UE capabilities
• LTE supports bandwidths of [1.25] [1.6] 2.5,
5, 10, 15 and 20MHz
• Two bandwidth capabilities for the UEs are
defined; 10MHz and 20Mz
• 10MHz capable UE supports all the
bandwidths up to 10MHz.
• 20MHz capable UE supports all the
bandwidths including 20MHz.
• Case of interest, 10MHz capability UEs
connecting to a 15/20MHz Node-B
20MHz UE (1.25, 1.6, 2.5, 5, 10, 15 and 20MHz)
10MHz UE (1.25, 1.6, 2.5,
5 and 10MHz)
20MHz
10MHz

75
Synchronization Channel
The SCH in the center of the 20 MHz band is used in initial cell search for all the UEs
and in neighbor cell search for 20 MHz capability UEs.
The SCHs in the center of the left/right 10 MHz band are used in neighbor cell search
for 10 MHz capability UEs.
For bandwidths 1.25, 1.6, 2.5, 5.0 and 10MHz bandwidth, a single 1.25MHz SCH is
transmitted at the center frequency.
Both hierarchical and non-hierarchical SCH structures are being considered for LTE.
SCH SCHSCH
RF carrier
Center of the
left useful band
Center of the
right useful band
1.25 MHz 1.25 MHz1.25 MHz
20 MHz BW

76
Broadcast Channel
The BCH is defined for 1.25 MHz and centered in the middle of the overall
transmission bandwidth.
BCH
10-MHz bandwidth
20-MHz bandwidth
5-MHz bandwidth
1.25-MHz bandwidth
2.5-MHz bandwidth

77
Cell Search
Cell site with 20-MHz transmission bandwidth
SCH
Center carrier frequency
Step 1:
Cell search using synchronization
channel
detect center 1.25 spectrum of
entire 20-MHz spectrum
Example: 10-MHz UE in 20-MHz cell site, SCH bandwidth = 1.25 MHz and BCH bandwidth = 1.25 MHz
Step 2:
BCH reception
BCH
BCH
reception
Initiate data transmission
using assigned spectrum
Step 3:
UE shifts to the center carrier frequency
assigned by the system and initiates
data transmission

78
Random Access Channel
• LTE supports both a non-synchronized random access and a synchronized random access.
• The non-synchronized access is used when the UE is not time synchronized. The non-synchronized access allows the Node B to estimate,
and, if needed, adjust the UE transmission timing to within a fraction of the cyclic prefix.
• The synchronized random access procedure may be used when the UE uplink is time synchronized by the Node B. The purpose of the
synchronized random access is to request resources for uplink data transmission. One of the objectives of using synchronized random
access procedure is to reduce the overall latency.
0.5 ms subframe
TRA-REP (10 ms radio frame)
TRA
Data transmissionBWRA
(Scheduled) Data transmission(Scheduled) Data transmission
Random- access preambleRandom- access preamble
Guard timeGuard time
Can be used for other random- access
channels or data transmission.

80
LTE Architecture
eNBs provide the evolved UTRA U-plane and C-plane protocol terminations towards the UE
The eNBs are interconnected with each other by means of the X2 interface. The eNBs
communicate with each other using X2 interface for support of handover etc.
The eNBs are also connected by means of the S1 interface to the EPC (Evolved Packet
Core). The S1 interface support a many-to-many relation between aGWs and eNBs.
eNB eNB
eNB
MME/UPE MME/UPE
S1
X2
X2
X2
EPC
E-UTRAN

81
LTE Protocol Architecture
eNBs provide the U-plane
(RLC/MAC/PHY) and C-plane
(RRC) protocol terminations
towards the UE.
The eNBs interface to the aGW via S1
eNB E-UTRA Node-B
aGW Access Gateway
RRM Radio Resource Management
RB Radio Bearer
RRC Radio Resource Control
RLC Radio Link Control
MAC Medium Access Control
PHY Physical Layer
SAE System Architecture Evolution
MME Mobility Management Entity
PDCP Packet Data Convergence Protocol

82
User-Plane Protocol Stack
The RLC and MAC sublayers (terminated in eNB on the network side) perform
functions such as scheduling, ARQ and Hybrid ARQ etc.
PDCP sublayer (terminated in aGW on the network side) performs functions such
as header compression, integrity protection and ciphering etc.

83
Control-Plane Protocol Stack
The RLC and MAC sublayers perform the same functions as for the U-plane.
RRC performs broadcast, paging, RRC connection management; RB control, mobility and UE
measurement reporting and control functions etc.
PDCP sublayer performs integrity protection and ciphering functions.
NAS (Non-Access Stratum) handles functions such as SAE bearer management; authentication, idle
mode mobility handling, paging origination, security control for the signalling between aGW-UE.

84
LTE Standard Development Schedule
LTE Study Item LTE Work Item
Kick-off
LTE Workshop
Nov. 2004
Dec. 2004 Sep. 2006 Sep. 2007
Jun. 2006 Jun. 2007
The LTE standard expected to be ready by September 2007.
A 3-month delay introduced in study item/work item completion.

Lte tutorial

More Related Content

What's hot (20)

Viewers also liked (8)

Similar to Lte tutorial (20)

Recently uploaded (20)

Lte tutorial