Int. J. Communications, Network and System Sciences, 2009, 2, 822-826
doi:10.4236/ijcns.2009.29095 Published Online December 2009 (http://www.SciRP.org/journal/ijcns/).
Copyright © 2009 SciRes. IJCNS
Adaptive Co-Channel Interference Suppression Technique
for Multi-User MIMO MC DS/CDMA Systems
Prabagarane Nagaradjane1, Arvind Sai Sarathi Vasan2,
Lakshmi Krishnan2, Anand Venkataswamy1
1Department of Electronics and Communication Engineering, SSN College of Engineering, SSN Institutions, Chennai, India
2Department of Electrical and Computer Engineering, University of Maryland, College Park, Maryland, USA
E-mail: prabagaranen@ssn.edu.in , {arvind88, lakshmik}@umd.edu, anandv267 @gmail.com
Received September 2, 2009; revised October 10, 2009; accepted November 16, 2009
Abstract
In this paper, an adaptive co-channel interference suppression technique for multi-user MIMO MC DS/CDMA
system is envisaged. MC DS/CDMA offers many advantages like flexibility, robustness, low PAPR and
spectral efficiency. In spite of these advantages, performance of MC DS/CDMA system is greatly impaired
by interference. Common interferences, which degrade the performance of the system, are MAI and CCI.
Mitigating these interferences can directly increase the capacity of the system. In this work, an adaptive
co-channel interference suppression technique based on single-stage and two-stage MMSE IC is considered
for multi-user MIMO MC DS/CDMA system. Simulation results show that, at low SNR two-stage MMSE
IC outperforms single-stage, while at high SNR, single-stage provides better BER performance. Based on
this, a selection criterion has been propounded for improved system performance as a whole in interference
limited environment. Also, adaptive selection criterion resulted in better error performance.
Keywords: CCI, MIMO, MC DS/CDMA, MMSE IC, ML Decoding
1. Introduction
Multi-carrier transmission has recently gained enormous
attention for providing high data rate communications on
both forward and reverse channels. Multi-carrier trans-
mission is realized through OFDM.OFDM performs well
in frequency selective channels [1]. One of the most prom-
ising single carrier transmission schemes is CDMA,
which is robust to noise [2]. Also in the recent past, multi-
input multi-output (MIMO) has proved to provide very
high capacity without any increase in bandwidth or
power [3]. Combining these techniques leads to MIMO
MC DS/CDMA system, which is expected to meet the
demands of future broadband (4G) wireless wide-area
networks. Though MIMO MC DS/CDMA possesses many
advantages, it still suffers from the traditional impair-
ments of conventional CDMA systems like MAI and
CCI [4–6]. Of late, myriad research concentration has
been on proposing techniques for mitigating MAI (Mul-
tiple Access Interference) and multi-path, but very little
work has been carried out on CCI (Co-Channel Interfer-
ence) combating techniques. In this paper, we investigate
the performance of single-stage and two-stage CCI can-
cellation techniques for a MIMO MC DS/CDMA system.
MIMO is realized by employing space-time block codes.
At the receiver, we have employed two-stage MMSE
(Minimum Mean Square Error) IC (Interference Cancel-
lation) with ML (Maximum Likelihood) decoding [4].
We have considered a multi-user MIMO MC DS/CDMA
system, with k asynchronous co-channel users in the up-
link. Each of the k asynchronous co-channel users is
equipped with NT transmit antennas and they communi-
cate to a single base station, equipped with NR receive
antennas. In this multi-user environment, k x NT interfer-
ing signals will be arriving at the base station. Conven-
tional interference mitigation technique from k-1 co-
channel users requires NT x ((k-1) +1) receive antennas,
so as to suppress the co-channel interferences. But by
employing STBC the same can be achieved by using NR
receive antennas such that NR k, which exploits the
spatial and temporal structure. Also in this work we con-
sider two co-channel users, each equipped with two
transmit antennas and the base station equipped with two
receive antennas. The performance of MMSE IC and ML
decoding algorithm over MC DS/CDMA system with
each MC DS/CDMA transmitter in turn carrying multi-
user data is assayed for both downlink and uplink. The
paper is organized as follows: Section 2 introduces the
N. PRABAGARANEET AL.823
system model we have considered throughout this work,
Section 3 describes the improved MMSE IC with ML
decoding algorithm, Section 4 describes the two-stage
MMSE IC for MIMO MC DS/CDMA system, Section 5
expounds the performance results and Section 6 provides
the conclusions.
2. System Models
Here the system model that we have considered com-
prises two asynchronous co-channel users in a MIMO
MC DS/CDMA system in which each transmitting ter-
minal is a MC DS/CDMA transmitter with MIMO sup-
port realized through Space Time Block Codes (spatial
diversity). The two co-channel users communicate with
two receive antennas at the receiver, which performs
interference cancellation and then detects the transmitted
signal of each user. Figure 1 shows the MIMO MC DS/
CDMA transmitter .The user data at each transmitting
terminal after appropriate constellation mapping is
multiplied with a spreading code and the spread sym-
bols of each data are multi-carrier modulated. Then
the sum of all the carriers of the kth user composes the
output of the kth user signal Sk(t). The total output S (t)
is the sum of all the user signals. After HPA (high
power amplification), the final signal is transmitted.
The transmitted signal corresponding to the nth data
symbol of the kth user is
tfj
cm
k
qm
M
m
Q
q
k
m
k
nm
enTqTtpCdtS
2
,
1
0
1
0
)()(  
(1)
where, k is the user number, m is the carrier number, q is
the chip number, Tc is chip duration and T is symbol du-
ration and equals to Q Tc. Q is the length of user specific
spreading code. dm
k is the data of mth sub-carrier and kth
user, c
m,q
k is the qth spreading code of mth carrier of
kth user, pm(t) is Root raised cosine pulse of mth carrier
and fm is the mth carrier frequency. When the total num-
ber of users is K, the total transmitted signal correspond-
ing to the nth data symbol is

1
0
2
,
1
0
1
0
)(
K
k
ftmjk
qm
M
m
Q
q
k
mn eCdtS
(2)
This modulated stream is then passed through a STBC
encoder which groups the symbols according to a spe-
cific STBC pattern (G2) and then transmits the symbols
through multiple transmit antennas. The received vector
at the first receive antenna for the transmitted symbols d1
and d2 is expressed as
21
11
21
11
*
1
*
2
21
2
1
n
n
h
h
dd
dd
r
r (3)
where, h11 and h21 denotes the channel fading and n11
and n21 represents the additive white Gaussian noise
Figure 1. MIMO MC DS/CDMA transmitter.
(AWGN). Let u1, u2 and v1, v2 represent the symbols
transmitted over two consecutive symbol durations by
the corresponding two co-channel asynchronous users.
As shown in Figure 2, let h11, h21, g11 and g21 represent
the channel fading between all the transmit antennas and
the first receive antenna at the base station respectively
and h12, h22, g12 and g22 represent the channel fading be-
tween all the transmit antennas and the second receive
antenna at the base station respectively. Now, the re-
ceived vector at the receive antenna Rx1 over the two
symbol time period is expressed as
1122111122111111 nvgvguhuhr 
(4)
12
*
121
*
211
*
121
*
21112nvgvguhuhr (5)
The received vector at the receiver antenna Rx2 is
given by
2122211222211221 nvgvguhuhr 
(6)
22
*
122
*
212
*
122
*
21222 nvgvguhuhr  (7)
3. Improved MMSE IC ML Decoding
The overall received signal from each of the receive an-
tenna is expressed as
T
TT RRr 21
(8)
The channel fading coefficient between each transmit-
ter and the receive antenna is rearranged for each co-
channel user to form a generic channel matrix. For de-
tecting the first co-channel user data, the corresponding
channel matrix is represented as
22
11
GH
GH
H (9)
and for the second user, it is given by
22
11
~
HG
HG
H (10)
The MMSE weight matrix is calculated based on the
generic channel matrix and the weight matrix value dif-
Copyright © 2009 SciRes. IJCNS
N. PRABAGARANE ET AL.
Copyright © 2009 SciRes. IJCNS
824
Figure 2. MIMO MC DS/CDMA with two co-channel asynchronous users and adaptive base station receiver system model.
fers for each co-channel user. For each of the two users
considered here in this work, the MMSE weight matrix is
computed from the expression [4,5]
22
11 12 2
ˆ
HH
swr vwr v

ˆ
0
(16)
4. Two Stage Interference Cancellation
4
2IHHM H
 and 4
2
~
~
~
IHHM H
 (11)
The two stage interference cancellation proceeds with
first decoding the data from the two terminals. Then,
assuming each decoded value is correct, the data corre-
sponding to the other user is estimated using Maximum
likelihood estimation. i.e. first assuming the first terminal
decoded data is correct based on single stage MMSE IC,
the second user data is estimated from the decoded data.
This is carried out by first calculating x1 and x2.
where, M is the weight matrix for the first user and
M
~
is the weight matrix for the second user. HH represent
the Hermitian transpose of the channel matrix and σ2 is
the noise variance. To suppress the interferences from
other co-channel users, the inverse of the weight matrix
is multiplied with the columns of the channel matrix
that represents the channel fading for a particular co-
channel user i.e.
11 1
ˆ
.
x
RHc
and 22 2
ˆ
.
1
1
1hMw
and 1
1
1
~
~
hMw
(12) 0
RHc (17)
where, is the data decoded, corresponding to the
first user. The second user data is estimated from the first
user decoded data by using the maximum likelihood es-
timation given by
0
ˆ
c
where,
h1=first column of H or
H
~
h2=second column of H or
H
~
The MMSE Interference cancellation receiver sup-
presses both co-channel interferences and noise compo-
nents, which means that the mean square error or vari-
ance between the transmitted symbols and the estimate is
reduced. The maximum likelihood (ML) detection is
used to detect the transmitted symbols for the corre-
sponding user. The ML decoding estimates the symbols
by determining the minimum Euclidian distance of all
possible transmitted symbols from the received constel-
lation [4], given by
0
2
0112
ˆ
ˆarg min{}
sS
2
2
s
xGsx Gs
 (18)
where, S takes all possible values in the signal constella-
tion. The reliability corresponding to the estimated data
is given by the expression
2
0110 220
ˆ
s
2
ˆ
x
GsxG s
  (19)
Similarly, the first user data based on single stage
MMSE IC is estimated by computing y1 and y2.
}
ˆˆ
{minarg
ˆ2
22
2
11
ˆ
urwurwU HH
Uu

(13)
111
ˆ
.
yRGs
1
and (20)
222
ˆ
.
yRGs 1
and where, 1
ˆ
s
is the data decoded corresponding to the
second user. The first user data is estimated from the
second user decoded data by using the expression [4]
}
ˆ
~
ˆ
~
{minarg
ˆ2
22
2
11
ˆ
vrwvrwV HH
Vv

(14)
where, and V represent the two co-channel users
estimated data. ,, and takes all possible
values of the users signal constellation. The reliability of
the decoded signals are computed by
U
ˆˆ
1
ˆ
u2
ˆ
u1
ˆ
v2
ˆ
v1
2
1112
ˆ
ˆarg min{}
cC
cyHcyH

2
2
c (21)
The corresponding value of reliability for the esti-
mated data is
22
01 122
ˆ
HH
cwr uwr u
2
1111 221
ˆ
cyHcy Hc
 2
ˆ
(22)
ˆ
(15)
N. PRABAGARANEET AL.825
0
The receiver computes the overall reliability for the
two users i.e. 00cs

 and 11cs1

 .The
decision is made on the sets of symbols computed by
comparing the two reliabilities. The comparison is
made as [4]
if (01
)
00
ˆˆˆ ˆ
(,)( ,)cscs
else
11
ˆˆˆ ˆ
(,)( ,)csc s (23)
The system illustrated in Figure 2, consists of a re-
ceiver with a switch between a single and double stage
MMSE IC unit, CSI (Channel State Information) and a
decision unit. When the channel is slowly varying, the
receiver detects the symbols based on single stage tech-
nique. When the channel variation is rapid, two stage
MMSE IC is employed to detect the symbols. At present,
perfect channel knowledge is assumed at the receiver.
The entire detection takes place based on the instantane-
ous SNR available at the receiver.
5. Performance Analysis
In this section, we present the performance of adaptive
co-channel suppression technique for a multi-user MIMO
MC DS/CDMA system. The simulation results for the
single stage and two stage interference cancellation tech-
niques are shown in Figures 3 and 4. The channel model
considered is quasi-static Rayleigh fading channel, which
is built on the classical understanding of Doppler shift
and delay spread. The modulation scheme employed is
BPSK, as it provides the best system throughput for
MIMO realization based on STBC. The number of channel
realizations considered for uplink and downlink is 5000
and 10000 respectively for each value for SNR. Table 1
summarizes the simulation parameters.
Simulation results divulge that, at low SNR value,
two stage interference cancellation techniques perform
well whereas at high SNR value single stage MMSE IC
with ML decoding provides better BER performance.
Hence, a trade off can be made in selecting the inter-
ference cancellation techniques at the receiver when the
SNR dwindles. This can result in better performance of
MIMO MC DS/CDMA system in an interference lim-
ited environment as switching of IC can be made in an
adaptive manner.
Figure 3 shows the uplink performance of MIMO MC
DS/CDMA system with single stage IC, two stage IC
and adaptive IC for two co-channel users. Each user data
is spread by a spreading factor of 32. Here in each MC
DS CDMA system one user is accommodated. It can be
discerned that adaptive IC outperforms both single stage
and double stage. The same performance can also be
realized over the downlink channel. Figure 4 elucidates
Table 1. Simulation parameters.
PARAMETERS VALUES
Spreading code Walsh Hadamard
Uplink 5000
Number of
channel reali-
zations Downlink 10000
FFT Size 128
Cyclic Prefix 1/8
Spreading Factor 16 or 32
Data Modulation BPSK
Channel Model Rayleigh
Number of Transmit antennas 2
Number of Receive antennas 2
Figure 3. Performance of adaptive co-channel interference
scheme for 2 co-channel users over an uplink communica-
tion channel for MIMO MC DS/CDMA system.
Figure 4. Performance of adaptive co-channel interference
scheme for 4 co-channel users over downlink communica-
tion channel for MIMO MC DS/CDMA system.
Copyright © 2009 SciRes. IJCNS
N. PRABAGARANE ET AL.
Copyright © 2009 SciRes. IJCNS
826
the performance of the same system with four co-channel
users, with each user spread by a spreading factor of 16
over downlink communication channel. Here also adap-
tive switching scheme provides better BER performance.
6. Conclusions
In this work, we considered a two stage MMSE co-channel
interference cancellation receiver for MIMO MC DS/
CDMA systems. MC DS /CDMA can be realized as a
prominent air interface for 4G Broadband communica-
tions; however, capacity of such systems is limited by
interference. Mitigating the various interferences can result
in confronting the future generation wireless networks
needs. In this paper we have analyzed a two stage IC
technique for a multi-user environment. Results of our
analysis reveal that a trade off could be made in selecting
the IC techniques for mitigating CCI. It could be dis-
cerned that at low SNR values two stage has resulted in
better performance because of its iterative nature while
at high SNR values, single stage performs better. Also,
it is expounded from our analysis that the adaptive in-
terference cancellation receiver has resulted in better
suppression of CCI.
7. References
[1] L. Hanzo, L-L. Yang, E-L. Kuna, and K. Yen, “Single and
multi-carrier DS-CDMA multi-user detection, space-time
spreading, synchronization and standards,” IEEE Press, 2003.
[2] M. K. Simon and M. S. Alouini, “BER performance of
multi-carrier DS-CDMA systems over generalized fading
channels,” IEEE International Conference, 1999.
[3] E. Bigieri, R. Calderbank, A. Constantinides, A. Gold-
smith, A. Paulraj, and H. Vincent Poor, “MIMO wireless
communications,” Cambridge University Press, 2007.
[4] Anand. V, Arvind. S, and Lakshmi Krishnan, “Investiga-
tions on the performance of MIMO assisted Multi Carrier
DS/CDMA system with multi-user detection for 4G mobile
communications,” Dissertation, SSN Institutions, 2009.
[5] Prabagarane Nagaradjane, Arvind Sai Sarathi Vasan, and
Lakshmi Krishnan, “A robust space time co-channel inter-
ference mitigation and detection technique for multi-user
MIMO multi-carrier DS/ CDMA systems,” Proceedings of
IEEE International Conference, Wireless Vitae, 2009.
[6] S. Kondo and L. B. Milstein, “Performance of multi-car-
rier DS CDMA systems”, IEEE Transactions on Com-
munications, Vol. 44, No. 2, February 1996.