A 4x4 MIMO-OFDM System with MRC in a Rayleigh Multipath Channel for WLAN

F.O. Tade and Y. Sun

University of Hertfordshire
Abstract: This paper evaluates a 4x4 MIMO-OFDM system using transmit diversity with MRC at the
receive end for WLAN applications. We compare different MIMO detection algorithms with different
modulation schemes. The combined MIMO-OFDM using the combined MMSE-SIC method for the
IEEE 802.11a/n standards is then implemented and the performances are given.
1 Introduction
The term MIMO is used to describe systems that employ the use of multiple antennas at both transmitter and
the receiver so as to improve performance by achieving higher bit error rates. It is one of the major
developments in the third generation wireless communication system and is internationally researched [1-8].
The signals are transmitted in multiple paths and therefore introduce spatial diversity on the data stream in the
channel. It is unlikely that all the paths would encounter severe fading at the same time which allows the MIMO
scheme to improve the signal liability in a natural wireless environment.
MIMO systems [3] have become attractive trends for broadband wireless communications such as wireless LAN
(IEEE 802.11n), WCDMA and WiMAX (IEEE 802.16); this is partly due to the significant increase in data
throughput and link range without the need to either increase the transmit power or the system bandwidth [8].
In this paper, we investigate a 4x4 MIMO system using different modulation techniques. A brief explanation of
the different system components is given. This MIMO system is then combined with an OFDM signal based
partially on the parameters set in the IEEE 802.11a and 802.11n standards. The results obtained using MATLAB
to simulate N bits in a Rayleigh multipath environment were discussed.
2 MIMO System and MIMO Detection
2.1 System Description
Figure 1 shows a typical MIMO system. To evaluate the MIMO system, a linear representation of its associated
signals is required. Assuming a system comprising N transmit and M receive antennas where
n
x represents the
transmitted signal,
mn
h is the relevant entry in the channel matrix, H and
n
v is the noise with a variance of
2
o ,
the received signal
m
y is represented as follows:
1,2,...M m
1
= + =

=
n
N
n
n mn m
v x h y
(1)
The multiple transmit and receive antennas create channel coefficients [7] that can be realised in the form of an
M x N channel matrix shown below.
(
(
(
(

=
MN M M
N
N
h h h
h h h
h h h
H
...

...
...
2 1
2 22 21
1 12 11

(2)
Therefore, at any given time, t, the received signal can be expressed in the matrix format as shown below.
) ( ) ( ) ( t n t Hx t y + =
(3)
where y(t), x(t) and n(t) are the vectors of y
m
, x
n
and v
n
, respectively.

Figure 1: A typical MIMO system

2.2 Minimum Mean Squared Error, MMSE Detector
This is an algorithm which performs better than the standard ZF detector [3] under noisy conditions. Although it
does not fully eliminate ISI like the ZF algorithm; it substantially reduces the total noise power experienced at
the receiver [4].
2 2 1
2 2 1
( ( / ) )
( ( / ) )
H H
n s MMSE
H
n s MMSE
y H H I H y
y H I H y
o o
o o

+
= +
= +
(4)
where
n
o and
s
o represent the noise power and received signal power respectively and I represents an identity
matrix. It should be noted that when
n
o o >>
s
, the MMSE estimate equates the ZF estimate.
2.3 Ordered Successive Interference Cancellation
This is a form of non-linear equalization and is the detection algorithm utilised in this paper. The major
impairment in MIMO systems is Co-Antenna Interference, CAI. This impairment is readily avoided using the
OSIC algorithm [6]. This algorithm recursively detects the incoming sub-streams i.e. layers. It would initially
detect the strongest layer; i.e. sub-stream with the highest SNR, and then subtracts it from the original received
signal to properly eliminate the CAI. This process repeats for subsequent layers; detected on a basis of their
signal strength; until all the available sub-streams / layers have been detected and subtracted from the original
received signal [8].
SIC systems need initial estimates to perform efficiently. These initial estimates are obtained from the output of
the MMSE detector. It can behave in two manners. Firstly, it can subtract the soft information from received
signal, which leads to little or non-existent error propagation but acquires an accumulating noise effect for weak
users. Secondly, it can subtract hard information from the received signal leading to little or no noise
accumulation but possible error propagation. Successive interference cancellation could be done in a circular
manner to improve the performance at the expense of low convergence and thus high complexity. MAI is
reduced and near/far problem increased. Cancelling the strongest signal has the most benefit and is the most
reliable cancellation [5]. Therefore being the most reliable cancellation, the SIC algorithm would mainly suffer
from error propagation. It also requires channel estimate at the receiver [5]. Pre-filtering for general SIC systems
normally adopt either a ZF or an MMSE detector [6].

TxN

Rayleigh
Fading
Channel
Tx1
Tx2
Rx1
Rx2
RxM

Transmitter Receiver
Data Estimated
Data
3 MIMO OFDM System and Implementation of IEEE 802.11 Standards
Orthogonal Frequency Division Multiplexing (OFDM) is one of the most popular modulation schemes available
due to its significant advantages over other schemes such as BPSK, QPSK and n-QAM. It is used in wireless
LAN standards like 802.11a/g/n, HIPERLAN/2, Digital Video Broadcasting standard (DVBT). The latest draft
of the new IEEE 802.11n standard proposes high improvements on older standards with respect to the increased
throughput experienced in the physical layer (PHY); with data rates up to 600 Mbps [8]. This is achieved
through the combination of OFDM with MIMO techniques.
An understanding of how to interpret the standards is required to fully understand the increased throughput in
the proposed newer drafts of each 802.11 standard. As an example, the IEEE 802.11a standard is shown in
Table 1 and examined further.
Parameter Value
FFT Size 64
Used Sub-Carriers 52
Cyclic Prefix n/4
Subcarrier spacing 312.5kHz
Index of used sub-carriers {-26 to -1, +1 to +26}
Data symbol 3.2us
Table 1: IEEE 802.11a standard
From Table 1, it is noted that the total number of subcarriers is 64 with a sampling frequency of 20MHz. In the
802.11n standard, operation at 40MHz is supported. Only 52 of these are used in the IEEE 802.11a standard but
56 are used in the 802.11n standard; these are the sub-carriers used to transmit the actual data sequence. With
the symbol duration at 3.2s, therefore the total spacing between sub-carriers is kHz kHz
T
f 5 . 312
1
= = A
. A total of 16 samples are required, therefore 0.8s is the duration of the cyclic prefix leading to total symbol
duration of 4s.
4 Simulation Results
We first studied the performance and comparison of different modulation schemes (BPSK, QPSK, 16PSK and
16QAM) and different detection methods (ZF and MMSE) for 4x4 MIMO with MRC. The results of this study
are presented in Figure 2.

(a) (b)
Figure 2: Comparison of different modulation schemes for 4x4 MIMO with MRC: (a) ZF and (b) MMSE
A MIMO-OFDM combination using MRC at the receive end was then investigated in a Rayleigh multipath
environment. We try to achieve maximum gain by sending copies of the same OFDM through a 4x4 MIMO
System. This is also called transmit diversity in OFDM systems. Maximum Ratio Combining was then used to
boost the received estimated signal. Different configurations of the parameters in the 802.11a and 802.11n
standards were investigated to examine the effect on the received BER values.
A data signal with N = 10
5
was modulated using OFDM for a 4x4 MIMO System. The received signal was then
passed through a MRC stage to boost the performance. This data signal was investigated using OFDM partially
based on the 802.11a/n standards and the results for 802.11a and 802.11n are shown in Figure 3(a) and (b)
respectively. The cyclic prefix adds redundancy to the system so therefore, an account of this occurrence has
been accounted for in the form of wasted energy.

(a) (b)
Figure 3: BER values for 4x4 MIMO-OFDM with MMSE/SIC/MRC: (a) IEEE 80211a and (b) IEEE8011n
5 Conclusions
The choice of modulation scheme depends on the scenario with respect to the factors of transmission required.
For instance, 16PSK and 16QAM transmit more symbols than QPSK and BPSK. The 4x4 MIMO System was
investigated with receiver architecture based on ZF and MMSE receiver. As expected, due to increasing the total
number of transmitted symbols per second, there would be an increase in the total number of errors and
therefore the need to operate at a higher SNR values. The combination of OFDM with MIMO shows a form of
flooring in the curve which signifies the need for additional coding to fully utilise the OFDM scheme. This
flooring is a result of both inter-subcarrier and inter-antenna interference.

References
[1] F. B. Gross, Smart Antennas for Wireless Communications with Matlab, McGraw-Hill, 2005
[2] G. J. Foschini and M. J. Gans, "On limits of wireless communications in a fading environment when using
multiple antennas", Wireless Personal Communications. January 1998.
[3] S. Lokya and A. Kouki, New compound upper bound on MIMO channel capacity, IEEE Communication
Letters, Vol. 6, No.3, 2002, pp 96-98
[4] P. Wolniansky, G. J. Foschini, G. D. Golden, and R. A. Valenzuela, VBLAST: An architecture for realizing very
high data rates over the rich scattering wireless channel, URSI Int. Symp. Signals, Syst., and Electronics, 1998
[5] R. Sinha, A. Yener and R.D. Yates; Non-coherent multi-user communications: multistage detection and selective
filtering, EU RASIP Journal on Applied Signal Processing, 2002:12, 14151426
[6] P. W. Wolniansky, G. J. Foschini, G. D. Golden, and R. A. Valenzuela, V-BLAST: An architecture for realizing
very high data rates over the rich-scattering wireless channel, ISSE, Pisa, Italy, Sept. 1998.
[7] S. Lokya, Channel capacity of MIMO architecture using the exponential correlation matrix, IEEE
Communication Letters, Vol. 5, No 9, 2001, pp 369-371
[8] A. Paulraj, R. Nabar and D. Gore, Introduction to Space-Time Wireless Communications, Cambridge Press, 2003.