UNIVERSITETI POLITEKNIK I TIRANËS

FAKULTETI I TEKNOLOGJISË SË INFORMACIONIT

DEPARTAMENTI I ELEKTRONIKËS DHE TELEKOMUNIKACIONIT

PROJEKT KURSI

Lënda: Sistemet e Avancuara të Telekomunikacioneve

TEMA: “Performance Evaluation of LTE Downlink with MIMO Techniques,

evaluation of Bit Error Rate and Throughput as a function of SNR”

Punoi: Armela Kasa Pranoi: Prof. Asoc. Olimpjon Shurdi

Gledi Karaj
Orjeta Gjini
Petrit Kallajxhiu
Rea Deda

Tiranë, 2022
 Abstract

Long Term Evolution (LTE) of the Universal Mobile Telecommunication System (UMTS), also known as the Evolved
Packet System (EPS), is a transient advancement in mobile communications. The constant growth in demand for high-speed
network connections, low latency and delay, low error rates, and resilience necessitates such a revolution, because current users
and network applications have grown increasingly reliant on these needs for efficient functioning and performance.
LTE when used with Multiple-Input, Multiple-Output (MIMO), meet the growing throughput requirements and system
stability for users. This is investigated by understanding the throughput of LTE downlink in both spatial multiplexing mode and
transmission diversity mode. Performance reductions in the LTE framework have been analyzed. All simulations are done in
MATLAB. Simulations include bit error estimation (BER) that is verified for signal-to-noise ratio (SNR).
There are several array configuration and polarization, transmission, and detection techniques that may be employed to
meet various functional and performance goals. There are single polarized arrays and cross polarized arrays (which can be
compact or disconnected); transmission strategies include diversity schemes, notably transmit diversity and spatial multiplexing.
The report evaluates the performance of LTE including MIMO in ways to improve LTE system capacity, average LTE
throughput in different groups and BER performance against SNR. The simulation results are shown in the last section.
1. Introduction
1.1. Background and general overview

The demand for high-speed and extensive network access in mobile communications grows by the day, as the number of
users grows and applications with larger network resource demands are regularly created. As a result of this trend, mobile
communications have seen substantial advancements over the previous two decades, owing to extensive study.
The 3GPP Long Term Evolution (LTE) system represents the evolutionary transition from third generation (UMTS) to
fourth generation mobile technology. The initial work on LTE began with the completion of feasibility studies in version 7 of the
3GPP UMTS standards. This version also offered more enhancements to High Speed Packet Access (HSPA).
LTE along with extensive deployment and optimization of MIMO has numerous effects on LTE to further enhance data
flow. However, some unique challenges arise from MIMO and this requires network measurement and optimization techniques.
These are also called next generation wireless networks. MIMO systems in their transmission use many antennas to transmit (Tx).
The signal is transmitted in transmission (Tx) to the receiver antenna (Rx) at the same frequency. Wireless networks already use
MIMO, which is commonly used in wireless networks. In next generation networks, MIMO technology is a standard feature and
plays an important role in increased data rates and overall system capacity. In this research, LTE which is 4G wireless
communication is taken as the reference framework. The LTE framework has been analyzed to reduce its performance.
Simulations are done using MATLAB. The design and simulation of the OFDM system is done with cyclic prefix. The Bit Error
Rate (BER) is checked by changing the SNR value. LTE together with MIMO provide added capacity when using standard
antenna technique.

1.2. Generations of mobile communication

Mobile communication has become more popular in recent years due to the rapid revolution in mobile technology. Mobile
phones typically use a wireless telecommunication system which was first introduced in the 1980s. Since then there has been a
development of Telecom Generations (G) to the present day 4G and 5G coming in development progress. 1G technology was a
network with only voice calling capability and the maximum speed was 2.4 Kbps. 2G enables transmission of data such as text
messages (SMS - Short Message Service), transfer or files (MMS - Multimedia Messaging Service), but not video. In 3G
technology the data transmission increased up to 2Mbits / s, which allows you to send or receive large email messages.4G is the
fourth generation of mobile telecommunication which appeared in 2008.It was based on LTE standards (Long Term Evolution)
and LTE Advanced and offer a range of communication services including video calling, real-time translation and video voice
mail.5G generation will be based on lower cost, lower battery consumption. There will be much faster data transmission, 5G
speed will be 1Gbit / s.

1.3. Advantages of mobile communication

Mobile communication has the following advantages:
 Flexibility: Wireless communication enables people to communicate with each other regardless of location. You do not
need to be in an office or on a telephone booth in order to pass and receive messages.
 Cost effectiveness: In wireless communication, there is no need for any physical infrastructure (wires or cables) or
maintenance practices. Therefore, the cost is reduced.
  Speed: Improvements can also be seen quickly. Network connectivity or accessibility was greatly improved in accuracy
and speed.
 Accessibility: With the help of wireless technology easy access to remote areas is possible.
 Continuous connectivity: Continuous connectivity ensures that people respond to emergencies relatively quickly.

2. LTE features, architecture

2.1. What exactly is LTE?
Long-Term Evolution (LTE) is a wireless broadband communication standard for mobile devices and data terminals based
on GSM/EDGE and UMTS/HSPA technologies in telecommunications. It improves on those standards' capacity and speed by
using a different radio interface and core network improvements. Because different nations use different LTE frequencies and
bands, only multi-band phones can use LTE in all countries where it is supported.  LTE frequencies and bands differ from country
to country, only multi-band phones can use LTE in all countries where it is supported. The World Telecommunication
Organization established the basic specifications. Other countries and businesses, on the other hand, play an essential part in LTE
design. The goal of LTE was to boost the capacity and speed of wireless data networks by employing new techniques and
modulations developed around the century. Another goal was to rethink and simplify the network architecture in an IP-based
system with substantially lower data transfer than the 3G system. Because the LTE wireless interface is incompatible with 2G and
3G networks, it must operate on its own frequency.

2.2. History of LTE

After the development of Wireless Local Access Networks (WLAN) and the widespread availability of mobile networks,
the demand for internet access began to grow everywhere. This is one of the reasons why mobile networks have been expanded to
allow not just voice telephony but also wireless internet access at high speeds and with a high degree of mobility. The user was
supposed to be able to carry on a phone call while on the bus or on the road.

Figure 1: GSM-2G cellular network architecture

Following the development of 2G, the third generation of mobile (3G) communication technologies was created with the
primary purpose of improving data capabilities and including higher-speed connectivity. While 2G already had a big number of
services and a well-developed architecture, 3G designers took the strategy of keeping the current core GSM mobile network intact
while adding extra mobile data functionality in parallel with Sonic Mobile's existing network.
 Figure 2: 3G system architecture

The next generation of mobile communication technologies, known as 4G or Long-Term Evolution, is what we use every
day (LTE). By 2014, the 4G rollout had progressed to the point illustrated in Figure 3, where lighter blue indicates that 4G has
been implemented and is in use, while darker blue indicates that 4G is still being discussed. The standard is developed by
the 3GPP (3rd Generation Partnership Project) and is specified in its Release 8 document series, with minor enhancements
described in Release 9. LTE is also called 3.95G and has been marketed as "4G LTE" and "Advanced 4G";[citation needed] but it
does not meet the technical criteria of a 4G wireless service.

Figure 3: LTE development worldwide

2.3. LTE network and infrastructure

LTE (Long Term Evolution) has developed out of UMTS. Most UMTS networks have already been upgraded to High
Speed Packet Access (HSPA) in order to boost packet data rate and capacity. HSPA stands for High-Speed Downlink Packet
Access (HSDPA) and High-Speed Uplink Packet Access (HSUPA). The goal is to increase HSPA-based radio network
performance in terms of spectrum efficiency, data peak rate, and latency, and to realize their full potential. WCDMA is a wireless
communication system based on five frequencies. Currently, WCDMA networks are deployed in UMTS band of around 2 GHz in
Europe and Asia, including Japan and America Korea.
LTE was the 4G successor to the 3G UMTS system, which was created to provide a more advanced mobile
telecommunications system. Around 2008, the plan was placed into its basic form, providing substantially faster data rates and
enhanced performance, as well as cheaper operating costs. Initial deployments offered modest improvement above 3G HSPA and
were dubbed 3.5G or 3.99G, but full LTE capability was quickly reached, allowing for full 4G performance. The earliest LTE
deployments were simply referred to as LTE, while later deployments were dubbed 4G LTE Advanced, and then 4G LTE Pro.
The radio access network was not only upgraded to 4G LTE, but the network architecture was also overhauled, resulting in
decreased latency and considerably higher performance.
Elements of LTE infrastructure include E-UTRAN (Evolved UMTS Terrestrial Radio Access Network), E Node B, and
EPC (Evolved Packet Core). Due to the saturation in 2G and 3G networks and quest for higher data rates, demand for global LTE
deployment is increasing considerably. With the evolution of the LTE network, service providers have been looking for data-only
networks, carrying voice in a packet data form.
One of the key hurdles for the growth of the LTE infrastructure industry is the lack of compatible devices, particularly for
LTE Advanced. Mobile phone manufacturers, on the other hand, are expected to release more devices that are compatible with
new LTE technology.
LTE Advanced is expected to fuel LTE infrastructure growth in the next years, as key operators around the world evaluate
and test LTE Advanced networks. In 2013, SK Telecom, based in South Korea, began rolling out LTE Advanced across the
country. Alcatel-Lucent SA, Antenova, CDG, Berkeley-Varitronics Systems, Cisco, Ericsson, Huawei Technologies Co. Ltd.,
Nokia Solutions and Networks B.V., Qualcomm Inc., and ZTE Corp. are all major competitors in this sector.
2.4 New LTE features
When compared to prior mobile networks, LTE has incorporated a number of new technologies. They allow LTE to
operate more efficiently in terms of spectrum utilization while simultaneously providing significantly higher data rates.
OFDM (Orthogonal Frequency Division Multiplex) was adopted for the LTE signal format because it allowed for the efficient
transmission of high-speed data bands while also giving a high level of resistance to reflection and interference. Because data was
sent to a wide number of carriers, the system was able to cope even if some were lost due to interference from reflections and
other factors. The access techniques utilized in the uplink and downlink were different: in the downlink, OFDMA (Orthogonal
Frequency Division Multiple Access) was used, and in the uplink, SC-FDMA (Single Carrier - Frequency Division Multiple
Access) was employed.
MIMO (Multiple Input Multiple Output): One of the main problems encountered by previous telecommunication systems
was that of multiple signals arising from the multiple reflections encountered. Using MIMO, these additional signal paths could
be used to advantage and were able to be used to increase throughput. It is vital to employ many antennas while utilizing MIMO
to discriminate between distinct pathways. As a result, designs utilizing 2 x 2, 4 x 2, or 4 x 4 antenna arrays are possible. Adding
additional antennas to a base station is relatively straightforward.
Downlink MIMO (Multiple Input Multiple Output), top-order uplink (16QAM) and downlink (64QAM) modulation, layer
2 protocol improvements, and continuous packet connectivity are all important aspects of HSPA version 7. These features can be
divided into two categories: data scale or capacity enhancement features vs internet browsing and power conservation features.
Up to four downlink transport operations (without MIMO) and operation of two uplink transporters are now conceivable, such as
the combination of 64QAM and MIMO. The downlink and uplink data rates are increased to theoretical maximum of 168 Mbps
and 23 Mbps, respectively.
SAE (System Architecture Evolution): Because 3G LTE requires extremely high data rates and minimal latency, it was
important to upgrade the system architecture to increase performance. Several functions that had previously been handled by the
core network were moved to the periphery. It was essentially a "nicer" form of network architecture. This reduces latency and
allows data to be routed more directly to its intended destination. EPC was created as part of the process of updating an Evolved
Packet Core to guarantee that packet data was treated as efficiently as possible.
Data over IP: 4G LTE is a data over IP system. In 3G UMTS, interrupted voice was included in the circuit, while LTE did
not allow for any interruptions. The initial plan was for carriers to provide data capability while OTT applications would provide
voice. While operators would lose a large amount of money, the item was a crucial revenue generator at the time. To address this,
the GSMA has established the Voice over LTE (VoLTE) scheme as the industry standard for voice connectivity. Due of the cost,
VoLTE necessitated the deployment of an IMS kernel, which slowed down the capabilities.

2.5 LTE Mobile Devices and LTE Uu Interface

The mobile device is referred to as the User Device in LTE specifications, as it is in UMTS specifications (EU). In
contrast to HSPA, which allows devices to support a wide choice of modulation and coding schemes as the standard matured over
time, all LTE EUs soon adopted 64-QAM (Quadrature Amplitude Modulation) for downlink and antenna diversity. For LTE-
Advanced, new device categories have been defined.
Only terminals 1-4 are required in the uplink direction, and 16-QAM support is slower but more stable. Class 5 devices,
on the other hand, must support 64-QAM. All mobile devices must support MIMO transmission in the downlink direction, apart
from EU category 1, which was never implemented in practice. Several data streams are delivered at the same carrier frequency
from numerous antennas on the base station to multiple antennas on the mobile device using this innovative transmission system.
The receiver can discriminate between distinct broadcasts and reconstitute the original data feeds if the signals arrive via different
paths. The number of data streams that can be delivered in simultaneously is determined by the number of transmitting and
receiving antennas. Most LTE networks and devices use 2 MIMO, which means two transmit antennas and two receive antennas.

CATEGORY 1 2 3 4 5
Maximum downlink data (carrier 20 MHz) 10 50 100 150 300
Maximum uplink data 5 25 50 50 75
Number of receiving antennas 2 2 2 2 4
Number of MIMO downlink feeds 1 2 2 2 4
Support for 64 QAM in the uplink direction No No No No Yes

Table 1: AE LTE categories

Depending on the geographical region, LTE networks are implemented in several frequency bands. Other radio
technologies, such as GSM and UMTS, are supported by most LTE-enabled devices. As a result, a typical LTE device now
supports not only seven or more LTE frequency bands, but also additional radio technologies. A device supplied in Europe
normally supports GSM bands 900 and 1800 MHz, UMTS bands 900 and 2100 MHz, and international GSM and UMTS roaming
bands 850 and 1900 MHz This is a challenge for antenna designers because a device's antenna sensitivity should be the same in
all non-roaming bands supported.

3. Radio Propagation Environment and Channel Modeling

3.1 Introduction
Because the physical channels and transmissions are sensitive to the characteristics of the radio environment, the radio
propagation environment is a significant factor in evaluating LTE performance. Realistic channel modeling is required for correct
evaluation of LTE link- and system-level performance, as well as for network design during the deployment phase.
Because the preferred spatial transmission mode and its performance are primarily dictated by it, standardized MIMO radio
channel models must be researched, and the correlation between the signals of the different antenna branches must also be
precisely predicted.
Furthermore, the spatial correlation features of the MIMO radio channel define the theoretical channel capacity's upper
limit. All the instantaneous space-time-frequency parameters that affect the configuration of diversity, beam shaping, and spatial
multiplexing techniques should be reflected in the applicable channel.

3.2 The Radio Propagation Environment

Dense urban, suburban, and rural areas are examples of radio propagation environments. Various items inside these
surroundings, such as buildings, trees, rocks, and people, have a positive or negative impact on signal transmission. This
circumstance generates a phenomenon known as fading, which has confronted communications engineers with a difficult
technological difficulty.

3.2.1 Fading
Fading is a term that refers to the time variation of received signal power caused by changes in the transmission medium
or path(s). Atmospheric conditions and the location of obstructions inside the propagation environment from the transmitter to the
receiver can induce fading. Multipath is the outcome of the latter of these two happening at the same time. As demonstrated in the
basic figure below, multipath occurs when a signal is reflected by barriers along the path, resulting in several copies of the signal
arriving to the receiver with various delays.

Reflective

Scattering

BS
Diffraction

multipath

Figure 4:Multipath propagation

3.2.2 Multipath Propagation

Diffraction, reflection, and scattering are all impacts of multipath. The apparent bending or deviation of waves around
obstacles, as well as the dispersion of waves beyond narrow gaps, is referred to as diffraction. When a radio wave travelling in
one media collides with another medium having different electromagnetic properties, reflection occurs. When a radio signal hits a
rough surface or a size significantly smaller than or on the order of the signal wavelength, it scatters.

3.3 Rayleigh Fading Channel model

This is a statistical channel model based on the assumption that the magnitude of a signal varies randomly or fades
according to a Rayleigh distribution as it goes through a communications channel. Multiple paths created by reflections from
objects in the radio channel can manifest in different ways in communication receivers depending on the degree of path difference
relative to the signaling rate, also relative to the wavelength of propagation, and the relative motion between the transmitter and
receiver. Due to central-limits type effects, which are based on the central-limit theorem, the channel impulse response will be
well-modeled as a Gaussian process (complex random variable) irrespective of the distribution of the individual components if
the occurrence of scatter is adequate, but if there is no dominant component to the scatter, such a process will have a zero mean
and phase evenly distributed between 0 and 2pi radians if there is no dominant component to the scatter. As a result, the channel
response will have a random amplitude and a Rayleigh distribution envelope.

3.4 MIMO Channel

Because MIMO systems use many antennas at the transmitter, receiver, or both, it's critical to think about the channel and
signal model in this system. The MIMO channel at any given time can be described as an MR x MT matrix for a system with MT
transmit antennas and MR receive antennas, assuming frequency-flat fading over the bandwidth of interest.

Figure 5:Mimmo Channel and matrix

For LTE, there are two types of MIMO channel models to choose from: Extended ITU models, such as Extended
pedestrian A, Extended Vehicular A, and Extended TU, are based on correlation matrices (EPA, EVA, and ETU). The 3GPP
Spatial Channel Model (SCM), Spatial Channel Model – Extension (SCME), and the IST-WINNER model are all geometry-
based channel models.

3.5 ITU Channel Model

The ITU channel models were primarily employed in the development of 3G radio access systems known as "IMT 2000."
Indoor office, outdoor-to-indoor, pedestrian, and automobile radio settings are among the major scenarios. Time delay spread and
its statistical variability, path-loss and shadow fading characteristics, multipath fading characteristics, and operational radio
frequency are the fundamental parameters for describing each propagation model. The International Telecommunication Union
(ITU) has proposed multipath channel models that are used to select essential multipath conditions in typical environments where
the average energy per bit to noise density ratio (Eb/No) or SNR requirements of various services for specific performance levels
are specified. Three different test environments have been proposed by the ITU in the ITU-R M.1225 Recommendation: 1) Indoor
office test environment; 2) Outdoor to indoor and pedestrian test environment; 3) Vehicular test environment.

3.6 Physical Channels

The LTE air interface comprises of physical channels and physical signals. The physical signals are created in Layer 1 and
they are used for system synchronization, cell identification, and radio channel estimation. The physical channels are used to
carry data form higher layer including control, scheduling, and user payload.

Physical Channel Function

Physical Downlink Shared Channel (PDSCH) User data and control information.
Physical Multicast Channel (PMCH) Data to multiple devices in a specific area. (such as
digital radio and mobile TV)
Physical Downlink Control Channel (PDCCH) Control messages.
Physical Broadcast Channel (PBCH) Broadcast information to all mobiles in a coverage
area.
Physical Control Format Indicator Channel Channel configuration information.
(PCFICH)
Physical Hybrid ARQ Indicator Channel Retransmission flow control status information.
(PHICH)
Physical Random-Access Channel (PRACH) Access request messages.
Physical Uplink Shared Channel (PUSCH) User data and control information.
Physical Uplink Control Channel (PUCCH) Control messages
Table 2:Physical Channels and their function

4. MIMO in LTE Analysis

4.1. LTE and MIMO
Data rates are limited by the available signal strength of the received signal's signal-to-noise ratio (SNR) in Shannon's
work on channel capacity. The bandwidth of the transmission and the data rates are related to the channel capacity. Any increase
in data rate necessitates a boost in the proportionality of the received signal's power. The data rate is equal to or greater than the
available bandwidth when bandwidth utilization is high.
The transmitter side uses multiple antennas, referred to as transmission diversity. The transmitted power will be in the
direction of the receiver because the approximation of the transmission diversity is based on the formation of the beam using
multiple transmitting antennas. Beam formation in transmission diversity has the potential to increase signal strength and allows
higher data rates. It is important to note that data rates are saturated beyond a certain point because transmitting and receiving
diversity can only work up to a certain point. To achieve high data rates, the transmitter and receiver ends are equipped with
multiple antennas.
Transmission bandwidth is another factor that affects the data rates that can be achieved in mobile communications. The
wider transmission band is supported by the provision of higher data levels. The effect of multi-path fading is on the most
significant challenge in broadband transmission. Using the pulse response of the radio channel, the transmitted signal is filtered.
A faded channel with many paths in the frequency field indicates the time that the channel frequency responses change.
The problem of time domain equalization is overcome by two approaches to broadband transmission, they are:
 Use of multi-carrier transmission schemes. Here the signal with the widest band is represented as the sum of the many
numbers of orthogonal narrowband signals. For example, OFDM transmission is a multi-carrier transmission used in
the LTE standard.
 A single carrier transmission scheme is used. This scheme benefits from the frequency of complexity which is the low
equation which is given by OFDM. This is done in the absence of its high fluctuations of transmitting power. An
example of this is in SC-FDM which is used as the technology in the LTE standard for uplink transmission.

MIMO methods are used to improve mobile communication in two ways. Both ways are:
  Increase overall data levels
 Increase reliability in communication links

4.2. PHYSICAL CHANNELS and LTE SIGNALS

In LTE, as defined in the 3GPP standards, previous dedicated channels are replaced by channels that are shared and there
is a reduction in the total number of physical channels. The MAC layer is connected to PHY by transport and the physical
channels are processed by the transmitter in PHY. In data transmission that includes down-link and up-link transmission, the
system uses DL-SCH (Down-link Shared Channel) and UL-SCH (Up-link Shared Channel) respectively as the transport channel
types. The physical channel carries the frequency sources for transmitting data to a specific transport channel. The corresponding
frequency channel is mapped to each transport channel. Further, in the physical channels and their respective transport channels,
there are some physical channels when the corresponding transport channels are missing. These are called control channels
indicated by L1/L2 which is used to discharge control information (DCI) and provide the required data. The relationship between
logical channels, physical channels, and transport in LTE has different downlink versus uplink transmissions.

4.3. MIMO

Multiple Inputs - Multiple Outputs or as it is otherwise known as MIMO is based on the idea of using multiple antennas
on the transmitter side and receiver side. The number of antennas varies from side to side or may be the same. The MIMO system
uses diversity techniques to improve overall system performance and can achieve significant system BER reduction.

Figure 6: Mimo logic

4.4. Spatial multiplexing

In this independent and complete data flow technique, data is transmitted simultaneously using each transmitting antenna.
Spatial multiplexing is used to transmit antenna ports for data. Simultaneously, the transmission of different modulated symbols is
possible through multiple antennas in the same sub-carrier on the frequency.
This means that spatial multiplexing has the ability to directly increase bandwidth efficiency and the resulting system has
high bandwidth utilization. Spatial multiplexing provides maximum benefits and advantages when transmitted using multiple
antennas that are not correlated. In this situation, the multi-way fading in communication links helps the actual performance.
Performance can be realized in spatial multiplication by solving a system of linear equations that describe the relationships
between receiving and transmitting antennas. The following figure illustrates a 2 x 2 antenna configuration for spatial
multiplexing.
 Figure 7: Spatial multiplexing for 2x2 antenna configuration in MIMO

5. Simulation and results

The simulation results will be presented and examined in this part, based on the simulated scenario stated below. The
performance of LTE Downlink using MIMO methods is investigated in terms of BER and Throughput for two key LTE channel
models, the flat-fading and ITU Pedestrian B channel. MATLAB is used to model and simulate the LTE downlink for various
MIMO approaches. So, we have one flat-fading channel and one multipath channel included in the scenario.
We are going to compare different MIMO technique: the standard SISO technique and Spatial Multiplexing (SM).
Different tests are done for 4 distinct modulation orders, QPSK, 16QAM, 64QAM and 128QAM. And at the receiver we will use
Soft Sphere and Zero and Forcing Detection.
At the table below are listed some parameters needed in the simulation:
Bandwidth 5MHz
Modulation QPSK, 16QAM,
64QAM, 128QAM
Cyclic prefix Cyclic prefix
IFFT size 512
Channel type Flat fading; ITU Pedestrian B

Receiver decoder type ZF, SSD

Channel coding Turbo
Number of iterations 1000
No. of Tx antenna 2
No. of Rx antenna 2
No. of users 10
Transmission mode SISO; Spatial Multiplexing (SM)
Table 3: parameters needed in the simulation

5.1. Performance Analysis and Comparison of SSD vs. ZF detection in Flat-Fading

5.1.1 QPSK
 The greatest throughput achievable is roughly 2.95Mb/s, as seen in Figure 8. In low SNR regions, SSD decoding beats ZF
decoding and this difference is especially noticeable in spatial multiplexing (SM), which requires additional SNR values for the
same throughput values up to 20 db. The graphs show that we get a smaller bit error rate with SSD, then ZF. In the low SNR area,
the SSD SISO follows ZF-SISO.

Throughput QPSK
3
2.5

Throughput [Mbps]
2
1.5
1
SSD SM ZF SM
0.5 SSD-SISO ZF-SISO
0 SSD TD ZF TD
0 5 10 15 20 25 30
SNR [dB]

Figure 8: Throughput Performance of SISO, TM and SM with ZF vs. SSD decoding in Flat-fading channel

In Fig. 9, it can be seen that bit error rates are lower with SSD detection than with ZF detection; a good illustration of this
is when we compare the SM curve with SSD (blue curve) and ZF (orange curve), there is obviously a large gap between them.
The BER curves of SISO with ZF and SSD, on the other hand, appear to be intertwined, making it impossible to discern which
has higher BER performance.

BER QPSK
0.181
SSD SM ZF SM
0.161
0.141 SSD-SISO ZF-SISO
0.121
SSD TD ZF TD
0.101
BER

0.081
0.061
0.041
0.021
0.001
0 2 4 6 8 10 12 14 16
SNR [dB]

Figure 9: BER Performance of SISO, SM with ZF vs. SSD decoding in Flat-fading channel

5.1.2. 16 QAM

For SISO and transmit diversity, SSD achieves nearly the same throughput as ZF, however when spatial multiplexing is
used, SSD achieves slightly higher throughput than ZF (SM). We can see this in Fig. 10. The maximum attainable throughput is
around 14.2Mb/s.
 Throghput 16 QAM
SSD SM ZF SM
16
SSD-SISO ZF-SISO
14

Throughput [Mbps]
12 SSD TD ZF TD
10
8
6
4
2
0
0 5 10 15 20 25 30
SNR [dB]

Figure 10: Throughput Performance of SISO, SM with ZF vs. SSD decoding in Flat-fading channel

According to Fig. 11, SM has the best BER for SSD receiver. SISO performance with detection techniques (ZF and SSD) has
similar BER with minor changes. The BER performance of SM with ZF detection was the worst as we can see in the figure.

BER 16 QAM
0.301 SSD SM ZF SM

0.251 SSD-SISO ZF-SISO

SSD TD ZF TD
0.201
0.151
BER

0.101
0.051
0.001
0 2 4 6 8 10 12 14 16 18 20
SNR [dB]

Figure 11: BER Performance of SISO, SM with ZF vs. SSD decoding in Flat-fading channel

5.1.3. 64 QAM

The performance shown in fig. 12 follows a similar pattern to that seen in 16 QAM, with the exception that throughput
values have grown as the order of modulation has increased, peaking at roughly 35Mbps with SM using SSD detection. For SISO,
the performance of SSD and ZF is relatively similar.

Throughput 64 QAM
40
35 SSD SM
Throughput [Mbps]

30 ZF SM
25 SSD-SISO
20
ZF-SISO
15
10 SSD TD
5 ZF TD
0
0 5 10 15 20 25 30
SNR [dB]

Figure 12: Throughput Performance of SISO, SM with ZF vs. SSD decoding in Flat-fading channel

The BER performance with 64 QAM is shown here; more SNR is needed for better performance. SISO technique with
SSD and ZF detection performs similarly. Best performance is archived from SSD SM with a BER of 10-3.
 BER 64 QAM
SSD SM ZF SM
0.351
SSD-SISO ZF-SISO
0.301
SSD TD ZF TD
0.251
0.201

BER
0.151
0.101
0.051
0.001
0 5 10 15 20 25 30
SNR [dB]

Figure 13: BER Performance of SISO technique with ZF vs. SSD decoding in Flat-fading channel

5.1.4. 128 QAM

The performance shown in fig. 14 follows a similar pattern to that seen in 64 QAM, with the exception that throughput
values have grown as the order of modulation has increased, peaking at roughly 41Mbps with SM using SSD detection. For SISO,
the performance of SSD and ZF is relatively similar.
As for the BER performance, it is also similar with 64 QAM. We still need more SNR in order to have a better
performance. Also, here SISO technique with SSD and ZF detection performs similarly. The worst performance is archived for
ZF detection for SM technique.

Throughput 128 QAM

45
SSD SM ZF SM
40
SSD-SISO ZF-SISO
35 SSD TD ZF TD
Throughput [Mbps]

30
25
20
15
10
5
0
0 5 10 15 20 25 30
SNR [dB]

Figure 14: Throughput Performance of SISO, SM with ZF vs. SSD decoding in Flat-fading channel
 BER 128 QAM
0.451
SSD SM ZF SM
0.401
ZF-SISO SSD-SISO
0.351
SSD TD ZF TD
0.301
0.251

BER
0.201
0.151
0.101
0.051
0.001
0 5 10 15 20 25 30
SNR [dB]

Figure 15: BER Performance of SISO technique with ZF vs. SSD decoding in Flat-fading channel

5.2. Performance and comparison of SSD vs ZF detection in ITU Pedestrian B Channel

5.2.1. QPSK

This section analyzes the throughput and BER performance of SISO and SM MIMO techniques for SSD and ZF detection
systems for an ITU pedestrian B channel. In general, detection by SSD gives superior results than detection by ZF, especially in
low SNR zones, according to the theory of MIMO detection approaches.
In the figure below, the spatial multiplexing curves with SSD and ZF detection (blue and red curves respectively) are
clearly visible at low SNR ranges between 0 and 5 dB, indicating that SSD achieves greater throughput. At 0dB, SM with ZF had
a throughput of 1.25Mbps, whereas SM with SSD had a throughput of roughly 2.6Mbps. 2.6Mbps is a significant improvement
over 1.35Mbps. The SISO technique keeps the throughput constant between SSD and ZF.

Throughput QPSK
3
Throughput [Mbps]

2.5
2
1.5
1
SSD SM ZF SM
0.5
SSD-SISO ZF-SISO
0 SSD TD ZF TD
0 5 10 15 20 25 30
SNR [dB]

Figure 16: Throughput Performance of SISO, SM with ZF vs. SSD decoding in Flat-fading channel

The BER curves are shown in Fig. 17, and the SSD SM scheme has very low error rates. Next is the SISO, where SSD
detection outperforms ZF detection. As seen in the figure SM and ZF detection performs worst at the range 0 and 4dB.
 BER QPSK
0.17 SSD SM ZF SM
0.15 SSD-SISO ZF-SISO
0.13
SSD TD ZF TD
0.11
0.09

BER
0.07
0.05
0.03
0.01
0 2 4 6 8 10 12
SNR [dB]

Figure 17: BER Performance of SISO technique with ZF vs. SSD decoding in Flat-fading channel

5.2.2. 16QAM

In 16 QAM, a pattern similar to that seen in QPSK is found. SISO achieves nearly the same throughput for both SSD and
ZF at 7dB. For SISO the detection algorithms have no influence on throughput, but there is an exception with SM, where
throughput performance is somewhat higher with SSD than with ZF between 5 and 15dB. Results shown in Fig. 18.

Throghput 16 QAM
16 SSD SM ZF SM
14 SSD-SISO ZF-SISO
Throughput [Mbps]

12 SSD TD ZF TD
10
8
6
4
2
0
0 5 10 15 20 25 30
SNR [dB]

Figure 18: Throughput Performance of SISO, SM with ZF vs. SSD decoding in Flat-fading channel

The BER performance of the ITU Pedestrian B channel is shown in Fig. 19, which shows that it has a lower bit error rates
than a flat-fading channel (fig.6.6). SM SDD has without a doubt the best bit error rate performance. SISO's performance was
better than ZF SM up to about 16dB SNR. Because all curves for each detection (ZF and SSD) for the two schemes are
comparable, there is no discernible influence of the detection techniques on SM and SISO. However, SSD is superior and
preferable for SM.

BER 16 QAM SSD SM ZF SM

0.301 SSD-SISO ZF-SISO
0.251 SSD TD ZF TD
0.201
0.151
BER

0.101
0.051
0.001
0 2 4 6 8 10 12 14 16
SNR [dB]

Figure 19: BER Performance of SISO technique with ZF vs. SSD decoding in Flat-fading channel

5.2.3. 64QAM
 Because low SNR values, such as those between 0 and 15dB, result in almost negligible throughput, this order of
modulation is ideally suited for regions with high SNR values, such as locations near the base station. For SISO, the performance
of both detection techniques is nearly identical. SM performs better with SSD detection than with ZF detection, suggesting that
raising the order of modulation does not improve one detection scheme's performance over the other.

Throughput 64 QAM
35 SSD SM ZF SM

Throughput [Mbps]
30 SSD-SISO ZF-SISO
25 SSD TD ZF TD
20
15
10
5
0
0 5 10 15 20 25 30
SNR [dB]

Figure 20: Throughput Performance of SISO, SM with ZF vs. SSD decoding in Flat-fading channel

The BER curves in Fig. 21 are quite similar to those produced in 16 QAM, but with a higher SNR necessary to achieve
low bit error rates. In SISO, SSD and ZF have nearly same performance. The poorest performance happens for SM with ZF
detection. As a result, SSD detection is preferred over ZF detection.

BER 64 QAM
SSD SM ZF SM
0.401
SSD-SISO ZF_SISO
0.301 SSD TD ZF TD
BER

0.201
0.101
0.001
0 5 10 15 20 25 30
SNR [dB]

Figure 21: BER Performance of SISO technique with ZF vs. SSD decoding in Flat-fading channel

5.2.4. 128QAM

Similar as in the 64QAM modulation, we get a better throughput with a bigger modulation order. For SISO, the
performance of both detection techniques is nearly identical. SM performs better with SSD detection than with ZF detection,
suggesting again as in the 64QAM modulation that raising the order of modulation does not improve one detection scheme's
performance over the other.

Throughput 128 QAM

45 SSD SM ZF SM
40 SSD-SISO ZF-SISO
Throughput [Mbps]

35 SSD TD ZF TD
30
25
20
15
10
5
0
0 5 10 15 20 25 30
SNR [dB]
 Figure 22: Throughput Performance of SISO, SM with ZF vs. SSD decoding in Flat-fading channel

The BER curves in Fig. 23 are quite similar to those in 64 QAM section, but with a better BER. In transmit diversity and
SISO, SSD and ZF detection have nearly same performance. In SM with ZF detection, is still the poorest. SSD detection is the
one with the best performance.

BER 128 QAM

0.401 SSD SM ZF SM
0.351 SSD-SISO ZF-SISO
0.301 SSD TD ZF TD
0.251
0.201

BER
0.151
0.101
0.051
0.001
0 5 10 15 20 25 30
SNR [dB]

Figure 23: BER Performance of SISO technique with ZF vs. SSD decoding in Flat-fading channel

Conclusions

The performance of LTE downlink using different MIMO approaches as Spatial Multiplexing (SM) in contrast to the
standard SISO system has been studied, analyzed, and evaluated in this paper. The performance is measured in terms of
throughput and BER, using different decoders (soft sphere - SSD and zero forcing -ZF decoders) used at the receiver in two
distinct channel models, flat-fading and ITU pedestrian B channel. For both technique SISO an SM, the flat-fading channel
performs better in both receivers for higher order modulation (16, 64 and 128 QAM) in the low SNR zones (up to 15 dB).
However, performance in the ITU pedestrian B channel is higher at low SNR, for low order modulation, such as QPSK in
our case. Performance for users that are far away from the base station is low in situations like ITU pedestrian B channel owing to
losses induced by the presence of many scattering, while performance for flat-fading channel is higher in these low SNR
locations, particularly when SSD is utilized.
The findings show that MIMO performs way better than SISO in both channel types, especially when SSD is used. At low
SNR locations, the flat-fading channel model outperforms the ITU pedestrian B channel when high order modulation is used.
Spatial multiplexing is suitable for obtaining extremely high peak rates.

References
[1] Gessese Kebede, Oladele Paul. (2010). Performance Evaluation of LTE Downlink with MIMO Techniques.MEE10:104.
[2] J. Zyren. “Overview of the 3GPP long term evolution physical layer” freescale.com, July 2007. [Online].
[3] S. Sesia, I. Toufik, M. Baker. LTE – The UMTS Long Term Evolution: From Theory to Practice. First Edition West Sussex: John Wiley & Sons, 2009.

Abstract

MATLAB Code
% Script for computing the BER for different modulation in a Rayleigh fading channel with 2 Tx,
% 2Rx MIMO channel Zero Forcing equalization
clear
N = 5*10^6; % number of bits or symbols
Eb_N0_dB = [0:25]; % multiple Eb/N0 values
nTx = 2; % Number of transmitters
nRx = 2; % Number of receivers
for ii = 1: length(Eb_N0_dB)
% Transmitter
ip = rand(1, N)>0.5; % generating 0,1 with equal probability
s = 2*ip-1; % modulation 0 -> -1; 1 -> 0
sMod = kron(s, ones(nRx,1)); %
sMod = reshape(sMod, [nRx,nTx,N/nTx]); % grouping in [nRx,nTx,N/NTx ] matrix
h = 1/sqrt(2) *[randn(nRx,nTx,N/nTx) + j*randn(nRx,nTx,N/nTx)]; % Rayleigh channel
n = 1/sqrt(2) *[randn(nRx,N/nTx) + j*randn(nRx,N/nTx)]; % white gaussian noise, 0dB variance
% Channel and noise Noise addition
y = squeeze(sum(h.*sMod,2)) + 10^(-Eb_N0_dB(ii)/20) *n;
% Receiver
% Forming the Zero Forcing equalization matrix W = inv(H^H*H) *H^H ------- ZF
% H^H*H is of dimension [nTx x nTx]. In this case [2 x 2]
% Inverse of a [2x2] matrix [a b; c d] = 1/(ad-bc)[d -b;-c a]
hCof = zeros(2,2,N/nTx) ;
hCof(1,1,:) = sum(h(:,2,:).*conj(h(:,2,:)),1); % d term
hCof(2,2,:) = sum(h(:,1,:).*conj(h(:,1,:)),1); % a term
hCof(2,1,:) = -sum(h(:,2,:).*conj(h(:,1,:)),1); % c term
hCof(1,2,:) = -sum(h(:,1,:).*conj(h(:,2,:)),1); % b term
hDen = ((hCof(1,1,:).*hCof(2,2,:)) - (hCof(1,2,:).*hCof(2,1,:))); % ad-bc term
hDen = reshape(kron(reshape(hDen,1,N/nTx),ones(2,2)),2,2,N/nTx); % formatting for division
hInv = hCof./hDen; % inv(H^H*H)
hMod = reshape(conj(h),nRx,N); % H^H operation
yMod = kron(y,ones(1,2)); % formatting the received symbol for equalization
yMod = sum(hMod.*yMod,1); % H^H * y
yMod = kron(reshape(yMod,2,N/nTx),ones(1,2)); % formatting
yHat = sum(reshape(hInv,2,N).*yMod,1); % inv(H^H*H)*H^H*y
% receiver - hard decision decoding
ipHat = real(yHat)>0;
% counting the errors
nErr(ii) = size(find([ip- ipHat]),2);
end
simBer = nErr/N; % simulated ber
EbN0Lin = 10.^(Eb_N0_dB/10);
theoryBer_nRx1 = 0.5.*(1-1*(1+1./EbN0Lin).^(-0.5));
p = 1/2 - 1/2*(1+1./EbN0Lin).^(-1/2);
theoryBerMRC_nRx2 = p.^2.*(1+2*(1-p));
close all
figure
semilogy(Eb_N0_dB,simBer,'mo-','LineWidth',2);
axis([0 25 10^-5 0.5])
grid on
legend('Flat-Fading channel (nTx=2, nRx=2)');
xlabel('SNR [dB]');
ylabel('Bit Error Rate');
title('BER');
================================================================
%% Configure MIMO Channel Object Using LTE MIMO Channel Object
% Configure an equivalent |MIMOChannel| System Object using the
% |LTEMIMOChannel| System Object. Then, verify that the channel output and
% the path gain output from the two objects are the same.
hMod = comm.PSKModulator;
modData = step(hMod,randi([0 hMod.ModulationOrder-1],2e3,1));
%%
% Split modulated data into two spatial streams.
channelInput = reshape(modData,[2 1e3]).';
%%
% Create an |LTEMIMOChannel| System object with a 2-by-2 antenna
% configuration and a medium correlation level.
hLTEChan = comm.LTEMIMOChannel(...
'Profile', 'EVA 5Hz',...
'AntennaConfiguration', '2x2',...
'CorrelationLevel', 'Medium',...
'AntennaSelection', 'Off',...
'RandomStream', 'mt19937ar with seed',...
'Seed', 10,... % No of users = seeds
'PathGainsOutputPort', true);
%%
% Filter the modulated data using the |LTEMIMOChannel| System object,
% |hLTEChan|.
[LTEChanOut,LTEPathGains] = step(hLTEChan,channelInput);
%%
% Create an equivalent |MIMOChannel| System object, |hMIMOChan|, using the
% properties of the |LTEMIMOChannel| System object, |hLTEChan|.
%
% The |KFactor|, |DirectPathDopplerShift| and |DirectPathInitialPhase|
% properties only exist for the |MIMOChannel| System object. All other
% |MIMOChannel| System object properties also exist for the
% |LTEMIMOChannel| System object; however, some properties are hidden and
% read-only.
hMIMOChan = comm.MIMOChannel(...
'SampleRate', hLTEChan.SampleRate,...
'PathDelays', hLTEChan.PathDelays,...
'AveragePathGains', hLTEChan.AveragePathGains,...
'NormalizePathGains', hLTEChan.NormalizePathGains,...
'FadingDistribution', hLTEChan.FadingDistribution,...
'MaximumDopplerShift', hLTEChan.MaximumDopplerShift,...
'DopplerSpectrum', hLTEChan.DopplerSpectrum,...
'SpatialCorrelation', hLTEChan.SpatialCorrelation,...
'TransmitCorrelationMatrix', hLTEChan.TransmitCorrelationMatrix,...
'ReceiveCorrelationMatrix', hLTEChan.ReceiveCorrelationMatrix,...
'AntennaSelection', hLTEChan.AntennaSelection,...
'NormalizeChannelOutputs', hLTEChan.NormalizeChannelOutputs,...
'RandomStream', hLTEChan.RandomStream,...
'Seed', hLTEChan.Seed,...
'PathGainsOutputPort', hLTEChan.PathGainsOutputPort);
%%
% Filter the modulated data using the equivalent |hMIMOChan| and use the
% |step| method to process data.
[MIMOChanOut, MIMOPathGains] = step(hMIMOChan,channelInput);
%%
% Verify that the channel output and the path gain output from the two
% objects are the same.
Throghput = isequal(LTEChanOut,MIMOChanOut)
BER = isequal(LTEPathGains,MIMOPathGains)