BER Performance Improvement for 4 X 4 MIMO Single Carrier FDMA System Using MMSE Equalization

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395 -0056
Volume: 03 Issue: 01 | Jan-2016 www.irjet.net p-ISSN: 2395-0072
© 2016, IRJET | Impact Factor value: 4.45 | ISO 9001:2008 Certified Journal | Page 1221
BER PERFORMANCE IMPROVEMENT FOR 4 X 4 MIMO SINGLE CARRIER
FDMA SYSTEM USING MMSE EQUALIZATION
Sharmila S1, Shanthi T2
1 PG student, ECE(VLSI Design), Kings College of Engineering, Tamil Nadu, India
2 Head of Department, ECE, Kings College of Engineering, Tamil Nadu, India
---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - The main objective of this project is to
design a soft decoding scheme to improve the Bit Error
Rate (BER) in the MIMO SC-FDMA uplink transmission.
The Long Term Evolution Advanced (LTE-A) uplink
transmission is covered as the main topic in this report.
This project mainly focuses on the uplink
transmission i.e., from Mobile to base station
transmission. The main focus is on the single-carrier
frequency division multiple access (SC-FDMA) scheme
which has been selected as the multiple-access scheme
of the LTE uplink. It also employs Multiple-Input
Multiple-Output (MIMO) because wireless
communication using MIMO links has emerged as one
of the most significant breakthroughs in modern
communications because of the huge capacity and
reliability gains promised even in worst fading
environment. The proposed system describes the basic
ideas of the MIMO SC-FDMA transmission systems and
focused and investigated the BER performance of the
Rayleigh wireless channel under 16 QAM ( Quadrature
amplitude modulation) modulation. Minimum Mean
Square Error (MMSE) equalization is performed in the
receiver side for better MIMO data detection. All
analysis was performed under ideal fading conditions
by the use of MATLAB which relates the SNR and the
error performance of MIMO SC-FDMA systems. All the
results obtained are simulated by using the MATLAB,
under Rayleigh channel conditions. The BER
performance of the proposed system is compared with
the Orthogonal Frequency Division Multiple Access
(OFDMA) MIMO systems which shows better results for
SC-FDMA. Thus the software implementation of this
MIMO SC-FDMA detector results in decreasing BER with
increasing SNR.
Key Words: SC-FDMA, MIMO, LTE-A, MMSE.
1. INTRODUCTION
LTE-A is a 4th generation mobile telecommunication
technology- 4G (International Mobile Telecommunications
Advanced (IMT Advanced) project. LTE-A was finalized by
the 3rd Generation Partnership Project (3GPP) in March
2011. LTE-A is not a completely new technology, rather it
is an enhancement to LTE. The main objectives of LTE-A is
to increase the peak data rate to 1 Gbps on the downlink
and 500 Mbps on the uplink, improve spectral efficiency
from a maximum of 16 bps/Hz in R8 to 30 bps/Hz in R10,
increase the number of simultaneously active subscribers,
and improve performance at cell edges. Many technologies
employed in LTE continue to be used in LTE-A, such as
orthogonal frequency division multiplexing (OFDM),
OFDMA, MIMO, and SC–FDMA. Some of the difficulties
encounter in the uplink of LTE-A are: 1) A problem
encountered in the design of receivers for LTE-A
communication systems is the detection of data from noisy
measurements of the transmitted signals. 2) Because of
OFDM’s high PAPR (peak to Average Power Ratio) and
related loss of efficiency, an alternative to OFDM was
desirable for the LTE uplink. 3) High rate of errors per bit.
4) Inefficient performance of modulation schemes such as
BPSK, QPSK, and 8-PSK.
The main goal of this project is to deliver data with less
BER for SC-FDMA LTE-A system. The bit error ratio (BER)
is the number of bit errors divided by the total number of
transferred bits during specified time interval. Some
solutions for the above problem are: 1) Instead of OFDMA,
SC-FDMA becomes a suitable scheme for the LTE uplink
because of its low PARP. Also MIMO technology gains
attention in wireless communications, because it offers
significant increases in data throughput and link range
without additional bandwidth or transmit power. 2) 16
QAM modulation scheme transmits 4 bits per symbol
efficiently. 3) The FDE equalization technique i.e., MMSE
decreases BER.
2. ARCHITECTURE OF MIMO SC-FDMA SYSTEM
2.1 TRANSMITTER
SC-FDMA transmitter will convert binary data into a
sequence of modulated subcarriers which is to be

transmitted through the radio channel. To do so many
signal process operations are required which as listed
below.
Fig -3: MIMO SC-FDMA Transmitter
Fig -4: MIMO SC-FDMA Receiver
First the data symbols are modulated by a base band
modulator to form a sequence of modulated complex
symbols. LTE uses an adaptive base band modulation
scheme so depending upon the channel it will adopt the
modulation formats. The common modulation being used
in LTE are Quadrature Phase Shift keying (QPSK), Binary
Phase shift Keying (BPSK), 16 level Quadrature amplitude
modulation (16-QAM) and 64-QAM. The next step is to
convert the serial modulated data into N parallel data
streams or group the data into blocks of N modulated
symbols.
Figure -5: 16-QAM Constellation
Then it performs an M point Discrete Fourier Transform
(DFT), this step will transform time domain modulated
symbols to frequency domain symbols. Nt ranges from 1 to
4. The next step is sub-carrier mapping which maps M DFT
output symbols to one of N orthogonal sub-carriers. N is
the number of orthogonal frequency sub-carriers, which is
greater than M, and M should be an integer multiple of N
so that N=M*Q where Q is the bandwidth expansion factor
or maximum number of users that can be supported by a
system. As an example, if M=64 and N=256 then Q=4 so
the maximum number of user that can be supported by the
system simultaneously are four. The result of sub-carrier
mapping is a set of complex symbols. There are two main
types of mapping schemes that a SC-FDMA system can
adopt, one is distributed and the other is localized.
1) Localized Scheme: In a localized scheme each user uses
a set of adjacent subcarriers to transmit its data. This
means for localized SC-FDMA (LFDMA) only a fraction of
the total bandwidth is used by one user. The advantage of
LFDMA is that it achieves multi user diversity in frequency
selective channel if each user is assigned subcarriers that
have high channel gain. The disadvantage of this scheme is
that it eliminates the chance of getting frequency diversity
in the channel. It also requires channel state information
(CSI) to map the data into the best adjacent symbols.
2) Distributed Scheme: In a distributed scheme the
subcarriers used by a user are spread over the entire
bandwidth. Since the information is spread it provides
inherent frequency diversity. One of the common versions
of distributed schemes is Interleaved SC-FDMA (IFDMA) in
which the subcarriers that are assigned to terminals are
equaled distant to each other. The disadvantage of this
scheme is the losing user diversity. An example of SC-
FDMA transmit symbols in the frequency domain is shown
in figure 6. Let consider N = 4 subcarriers per user, Q = 3
users, and M = 12 subcarriers in the system. where
Xl,Distributed = transmit symbols for distributed
subcarrier mapping scheme.
Xl,Localized = transmit symbols for localized
subcarrier mapping scheme.
Fig. -6: Distributed and Localized sub-carrier mapping

Instead of IDFT, Inverse FFT (IFFT) is performed which
transforms complex frequency domain symbols to time
domain signal more efficiently. Then each symbol is
modulated by a single high frequency carrier and
transmitted sequentially. If M=N then one can skip the
DFT, frequency Mapping and IFFT blocks and direct
modulated our time domain complex symbols to a single
frequency.
After that the Cyclic Prefix (CP), also called guard interval,
will be added. The CP means prefixing symbols with a
copy of the end of the symbol as shown in figure 7. The CP
has two purposes here, the first is that it is use as a guard
interval to eliminate the Inter Symbol Interference (ISI)
from the previous symbol. The second is that prefixing the
symbols with repetition of the end makes symbol periodic
and linear convolution with channel will changes to
circular convolution.
Fig -7: CP addition to data subcarriers
In frequency domain it is equivalent to point-wise
multiplication of symbols to channel frequency response.
To ease equalization, the length of CP should be minimum
equal to maximum delay in the channel or in other words
equal to the delay spread of the channel. Before
modulating the signal with high frequency to transmit
there is a pulse shaping filter that will shape the signal to
get the desired spectrum.
2.2 RECEIVER
At the receiver side exactly the inverse of transmitter data
transmission is performed. First demodulation of signal is
performed after that removal of CP takes place since CP is
an overhead it should be as optimized as possible. After
removing the CP the receiver transform the received
signal into frequency domain with the help of FFT. Nr
ranges from 1 to 4. It then de-maps the sub carriers and
then performs frequency domain equalization.
Normally the most common equalizer used is minimum
mean square error (MMSE) frequency domain equalizer.
The equalized signal is then transformed to time domain
by IDFT and detection is done in time domain.
2.2 MIMO DETECTION
A MIMO detector which is used for detecting receive
symbols, corresponds to symbols transmitted through
transmit antennas from receive signals, when the transmit
data transmitted by the terminal group are received
through receive antennas. In MIMO detection the detector
calculates an estimate of the transmitted signal as an
output of the detector based on the received signal and the
estimated channel matrix. After estimating and calculating
the channel matrix, the LTE-A system recovers the
transmitted signal from the received signal as an output of
the detector. Modern silicon technology allows detectors
to take full advantage of constrained maximum likelihood
(ML) search techniques, and can calculate a posteriori
probability (APP) soft information based on the received
symbols. ML search is the optimum detection method and
minimizes the bit error rate (BER). In general, the
detection algorithm can be classified into three types:
 Linear equalization algorithms
 Non linear equalization algorithms
 Optimal detection algorithms
Linear equalization algorithms include zero forcing (ZF)
and MMSE algorithms. Of these, ZF is the simplest
detection algorithm with the lowest computational
complexity. MMSE is a high complexity algorithm, but
offers high performance. Optimal detection algorithms
include maximum likelihood (ML) and sphere decoding.
These have preferable performance, but have the highest
complexity. Non linear equalization algorithms include
Successive interference cancellation (SIC), Parallel
interference cancellation (PIC),Vertical Bell labs layered
space time (V-BLAST) and QR decomposition algorithms.
They have lower complexity and better performance than
the optimum detection algorithms. The ZF receiver
completely nulls out the influence of the interference
signals coming from other transmit antennas and detects
every data stream separately. The disadvantage of this
receiver is that due to cancelling the influence of the
signals from other transmit antennas, the additive noise
may be strongly increased and thus the performance may
degrade heavily. Thus among these algorithms linear
equalization algorithm i.e. MMSE provides high
performance and hence MMSE algorithm is used in this
LTE-A system.
2.4 MMSE EQUALIZATION
MMSE equalization in receiver optimizes the noise and
offers a compromise between residual interference
between input signals and noise enhancement. There are
three kinds of equalizers: time domain equalizers,
frequency domain equalizers and combined equalizers.
For high ISI channel, time domain equalizers have huge
complexity and become unattractive. Among frequency

domain equalizers, zero-forcing and MMSE equalizer are
simplest ones.
MMSE equalizer has better performance than zero forcing
equalizer and is easy to implement. There are some
equalizers belong to the third type. MMSE equalizer is a
type of equalizer operating in both time and frequency
domains. This equalizer can achieve a better BER
performance with much higher algorithmic complexity.
Because of its simple architecture and relatively good
performance, frequency domain MMSE equalizer is chosen
in our implementation. MMSE equalizer minimizes the
mean square error between its output and the symbols
transmitted from the transmitter. Also MMSE equalizer
effectively minimizes the bit error rate compared to other
types of equalizers.
3. MATHEMATICAL MODEL
3.1 MATHEMATICAL ANALYSIS OF MIMO SC-FDMA
DATA TRANSMISSION
The data stream on each transmit antenna is grouped into
blocks of M symbols, as follows:
Sk
Nt(t)= [Sk
Nt(0), Sk
Nt(1),… Sk
Nt(M-1) ]T
Where T = represents the transpose operation, Nt = the
antenna index, M = DFT size, and Sk
Nt(t) = represents the
data on the transmit antenna for user, whose elements are
chosen from quadrature amplitude modulation
(QAM)constellation. Using the localized sub-carrier
allocation scheme, the DFT outputs are mapped to M sub-
carriers allocated to each user to produce Dk
Nt(t). The
localized sub-carrier mapping matrix for user is denoted
by
Tk
N,M= [0M*(K-1)M, IM, 0M*(N-Km) ] T
Where IM is an M-dimensional identity matrix. The
resulting FD SC-FDMA signal, Dk
nt(t), is transformed into
the time-domain (TD) through an -point inverse fast
Fourier transform (IFFT) operation, resulting in the TD
signals as follows.
Dk
Nt(t)= F-1
N Tk
N,M FM Sk
Nt(t)
Where FM is the normalized M-point DFT matrix F-
1
N is the normalized N-point IFFT matrix. Finally, a cyclic
prefix (CP) is inserted and the final SC-FDMA signal is
ready for transmission.
After the CP removal on antenna Nr at the SC-
FDMA receiver with receive antenna, the received signal is
denoted as
RNr =  
tM
nt
K
k 11
hk
Nr,nt  Dk
Nt(t)+ wnr
where
 = N -point circular convolution,
WNr represents the additive white Gaussian noise (AWGN)
on antenna ,
hk
Nr,Nt is the channel impulse response (CIR) between the
transmit antenna and the receive antenna for user . The
channel response can be formulated as follows.
H=












44434241
34333231
24232221
14131211
hhhh
hhhh
hhhh
hhhh
Hk
Nr,Nt= 


1
0
N
n
hk
Nr,Nt[n] e-2πkn/N
hk
Nr,Nt= 1/M [ 


1
0
M
n
Hk
Nr,Nt[k] ]e-2πkn/N
Using an N–point fast Fourier transform (FFT) and
performing the sub-carrier de-mapping, the FD signal of
user , received at antenna nr is denoted as
Yk
Nr= [Tk
N,M ] T FN rNr
The received signals are combined using the
Minimum Mean Square Equalizer. The equalized signal is
then transformed to time domain by IDFT and detection is
done in time domain. Multiple couples of data such as
𝑥0,𝑥1,…,𝑥3 are transmitted through different antenna
respectively and the received signals are 𝑦1,𝑦2,…,𝑦4. Then,
the relationship between transmitted and the received
data through the antenna configuration is as follows:
𝑦1=ℎ11 𝑥0+ℎ12 𝑥1+ℎ13 𝑥2+ℎ14 𝑥3+𝑧1
𝑦2=ℎ21 𝑥0+ℎ22 𝑥1+ℎ23 𝑥2+ℎ24 𝑥3+𝑧2
𝑦3=ℎ31 𝑥0+ℎ32 𝑥1+ℎ33 𝑥2+ℎ34 𝑥3+𝑧3
𝑦4=ℎ41 𝑥0+ℎ42 𝑥1+ℎ43 𝑥2+ℎ44 𝑥3+𝑧4
MMSE equalization is performed at the receiver
by
s=arg min||Hs-Y||2 +||σs||2
where σs = Inverse SNR of the transmitted signal
Fig -8: Signal detection in QAM constellation
Consider the first digit of the constellation point, 0
corresponds to 8 points in the left half plane and 1
corresponds to another 8 points in the right half plane
(Figure 8), detection steps for first digit is as follows: The
first digit of s can be estimated by finding the minimum
distance between s and 8 points of the left half plane and 8
points of the right half plane respectively. For the case
|Re{s}< 2|, which means it is in the square area marked
with dashes, the minimum distances between s and the 8
^

points in the left half plane should be the distance between
s and the 4 points which is the second column of the 8
points, these 4 points are denoted by α2 . Similarly the
minimum distances between s and the 8 points in the right
half plane should be the distance between s and the 4
points which is the first column of the 8 points in the right
half plane, these 4 points are denoted by α3. The 2nd, 3rd
and 4th digit can be estimate in the similar way.
4. RESULT AND DISCUSSION
This chapter describes the software requirements,
parameters and simulation procedures to evaluate the
BER performance in MATLAB R2011a. Table -1:
Simulation parameters
PARAMETER VALUE
DFT Size 512
Number of symbols 128
Signal constellation QAM
Bandwidth 20 MHz
Guard interval length 16
SNR 0 to 30 db
Subcarrier mapping Localized mode
Antenna configuration MIMO
Number of base station
antennas
4
Number of UE antennas 4
Equalizer MMSE-FDE
In this section, the simulation results are presented for
error rate analysis, which helps to characterize the
behavior of different fading channels. The SNR and BER
can be calculated as follows.
SNR= 10log10 [signal power/ noise]
BER = no. of errors / total number of bits send
Fig -9: simulation flow chart
The BER and SNR performance of the proposed and the
base (existing) architecture for a 4x4 MIMO detector, with
one resource block allocated to each user was evaluated
through MATLAB simulations.
0 5 10 15 20 25
0
0.01
0.02
0.03
0.04
0.05
0.06
SNR in DB
ERRORRATE
SYMBOL ERROR RATE vs SNR in DB
BASE
PROP
Fig -10: Comparison of base and proposed detector

0 5 10 15 20 25 30
10
-300
10
-250
10
-200
10
-150
10
-100
10
-50
10
0
10
50
BitErrorRate
SNR, dB vs Bit Error Rate
Fig -11: SNR vs. Error rate
Figure 11 show decreasing BER with increased SNR –as
would be expected i.e. at 5db the proposed (blue line)
method shows improvement than the existing (red line)
method. It indicates that at SNR~ 10db, the error rate of
base paper is 0.024 and the error rate of proposed work is
0.02. Likewise at SNR~ 15db, the error rate of base paper
is 0.036 and the error rate of proposed work is 0.03. The
results clearly demonstrate that for 16 QAM modulation
the MMSE detector are more efficient for MIMO SC-FDMA
transmission systems.
1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2
10
1.692
10
1.694
10
1.696
10
1.698
10
1.7
SNR, dB
AVERAGEMEANSQUAREDERROR
SNR, dB vs Bit Error Rate
Theory
Fig -12: SNR vs. Average Mean Squared Error
The BER v/s SNR curves are plotted in logarithmic vertical
scale by setting SNR along the x-axis and BER (or error
probability) along the y-axis. The figure 12 shows that the
average mean square error decreases linearly with the
increasing SNR.
5. CONCLUSION
This project gives a detailed overview of an MIMO SC-
FDMA system and compared the data transmission
accuracy with the OFDMA system and with frequency
domain equalization (FDE) system. The ultimate goal of
this project was fulfilled by proposing an entire work flow
to optimize the BER performance of the required
baseband signal processing. This work flow consisted
mainly of three steps: a literature study, research and
analysis of MIMO data transmission, and evaluation of a
low BER MIMO SC-FDMA system using MATLAB R2011a. A
soft decoding MIMO detector with reasonable complexity
was implemented for a MIMO SC-FDMA coded system,
resulting in a significant enhancement in the BER
performance. The BER and error probability is analyzed
by varying the SNR. The BER performance of the proposed
detection scheme is close to ML detection while the
reduction in the complexity is significant in large
constellation sizes. The above results clearly demonstrate
that the SNR values are found to be lower for MIMO
detectors compared to other detector systems.
REFERENCES
[1] Bai W., Yan D., Xiao Y, and Li S. (2009), “Performance
evaluation of MIMO SC-FDMA system with FDE
receiver,” in Proc. Int. Conf. WirelessCommun.
SignalProcess.(WCSP),pp.1–5.
[2] Dhivagar B., Kuchi K. and Giridhar K. (2013), “An
iterative MIMO-DFE receiver with MLD for uplink SC-
FDMA,” in Proc. Natl. Conf. Commun.(NCC),pp.1–4.
[3] Liu X., He X., Ren W. and Li S. (2010), “Evaluation of
near MLD algorithmsin MIMO SC-FDMA system,” in
Proc. Int. Conf. Wireless Commun.Netw.
MobileComput(WiCOM),pp.1–4.
[4] Lim S., Kwon T., Lee J. and Hong D. (2010), “A new
grouping-ML detectorwith low complexity for SC-
FDMA systems,” in Proc. IEEE Int. Conf.Commun.
(ICC),pp.1–5.
[5] Muhammed, M., Khalid, K., and Muhammed, U.,
(2012), “LTE-Advanced: Requirements and
Challenges for 4G Cellular Network”, Journal of
Emerging Trends in Computing and Information
Science, 665-670.
[6] Neshatpour k. Mahdavi M. and Shabany M, (2012), “A
low-complexityhigh- throughput asic for the sc-
fdmamimo detectors,” in Proc. IEEEInt. Symp.
(ISCAS),pp.3065–3068.
[7] Pan Z., Wu G., Fang S. and Lin D. (2010), “Practical
soft-SIC detection for MIMO SC-FDMA system with
co-channel interference,” in
Proc.Int.Conf.WirelessCommun.SignalProcess.(WCS),
pp.1–5.
[8] Pravin, W., Raut, S.L., Badjate, (2013), “MIMO- Future
Wireless Communication”, International Journal of

Innovative Technology and Exploring
Engineering,102-104.
[9] Studer C., Fateh S. and Seethaler D. (2011), “ASIC
implementation of soft-input soft-output MIMO
detection using MMSE parallel interference
cancellation,” IEEE Journal, Solid-State Circuits, vol.
46, no. 7, pp.1754–1765.
[10] Wang G., Yin B., Amiri K., Sun Y., Wu M. and Cavallaro
J R. (2009), “FPGA prototyping of a high data rate LTE
uplink baseband receiver,” in Conf.Rec. Asilomar
Conf. Signals, Syst., Comput., pp. 248–252.

BER Performance Improvement for 4 X 4 MIMO Single Carrier FDMA System Using MMSE Equalization

More Related Content

What's hot (20)

Similar to BER Performance Improvement for 4 X 4 MIMO Single Carrier FDMA System Using MMSE Equalization (20)

More from IRJET Journal (20)

Recently uploaded (20)

BER Performance Improvement for 4 X 4 MIMO Single Carrier FDMA System Using MMSE Equalization