Sei sulla pagina 1di 5

Electronic copy available at: http://ssrn.

com/abstract=1955930
International Conference on Information and Communication Technology (IICT-07), July 26-28, 2007
A Novel Interleaver for Interleave-Division Multiple-Access Scheme

M. Shukla*, N. Anil Kumar**, V.K. Srivastava**, S. Tiwari**

*Department of Electronics Engineering,
Harcourt Butler Technological Institute, Kanpur
**Department of Electronics & Communication Engineering,
Motilal Nehru National Institute of Technology, Allahabad

Abstract-In this paper, we suggested a practical interleaver for interleave-division multiple-access (IDMA) systems. In
IDMA scheme, the users are distinguished by different chip-level interleaving methods instead of by different
signatures as in conventional code-division multiple-access (CDMA) scheme. A set of interleavers is considered to be
practical if it has zero-correlation with each other and also easy to generate. A properly defined correlation between
interleavers can be used to formulate a collision criterion, where zero-correlation (i.e., orthogonality) implies no
collision. Simulation results show that the proposed interleaver performs similar to random interleavers as reported
in [2,14] IDMA system.

Index TermsIDMA, CDMA, orthogonal interleavers, correlation between interleavers, BER

I. INTRODUCTION

The Code Division Multiple Access (CDMA) has made its depth
in wireless communication very deeply. It offers well-known
features such as dynamic channel sharing, soft capacity, reuse
factor of one, low drop-out rate and large coverage (due to soft
handover), ease of cellular planning, robustness to channel
impairments and immunity against interference, etc. The
performance of conventional CDMA systems [1] is limited by
multiple access interference (MAI), as well as intersymbol
interference (ISI). Also, the complexity of CDMA multiuser
detection has always been a serious concern. The problem can
be seen from the angle of computational cost as well complexity
of multiuser detection algorithms.
The multiple access scheme recently studied in [1,2] introduces
the concept of interleaving as the only means for user
separation, and hence it is referred to as interleave-division
multiple-access (IDMA). IDMA not only inherits many
advantages from conventional code-division multiple-access
(CDMA), which includes robustness against fading and
mitigation of cross-cell interference, but also allows a very
simple chip-by-chip (CBC) iterative multiuser detection (MUD)
strategy while achieving impressive performance [1, 2]. In
multipath channels, the per-user complexity of the IDMA MUD
algorithm [2] is found to be O(L), which is independent of the
user number K and linear with the channel tap number L. It is
much lower than that of other alternatives, such as O(2
K
) for the
maximum a posteriori (MAP)-based method [3] and O(K
2
) for
the well-known minimum mean squared error (MMSE) method
[4].
In [5], an IDMA system that uses randomly and independently
generated interleavers is presented. With these interleavers, it
has been shown in [5] that IDMA system performs similarly and
even better than a comparable CDMA system.
In theory, the user-specific interleavers can be generated
independently and randomly. If this is the case, the base station
(BS) has to use a considerable amount of memory to store these
interleavers, which may cause serious concern when the number
of users is large. Also, during the initial link setting-up phase,
there should be messages passing between the BS and mobile
stations (MSs) to inform each other about their interleavers.
Extra bandwidth resource will be consumed for this purpose if
the interleavers used by the BS and MSs are long and randomly
generated. In this paper, we examine a power-interleaver
generation method to alleviate this concern. With this method,
the interleaver assignment scheme is simplified and memory
cost is greatly reduced without sacrificing performance.


Figure 1. IDMA Transmitter and Receiver



843
Electronic copy available at: http://ssrn.com/abstract=1955930
International Conference on Information and Communication Technology (IICT-07), July 26-28, 2007
In this paper, we are proposing a new interleaver which is a
combination of cyclic interleaver and diagonal interleaver. In
section II, there is brief description of uncoded IDMA, while in
section III; the brief description about correlation of interleavers
is given. In section IV, the design of Proposed Interleaver is
proposed. The computer simulations are given in section V. The
sections VI covers conclusion.



II. IDMA SCHEME


The IDMA scheme is shown in fig. 1 with all the necessary blocks,
however for simulation of IDMA with new interleaver, we have
considered uncoded IDMA scheme. The upper part of Fig. 1 shows
the transmitter structure of an IDMA system.

The data input properly modulated with BPSK modulation is
followed by a chip level interleaver that maps
( ) m
+
( ) K
d to
( ) K
x ={
( ) K
x , j=1, 2, ..J}.

We follow the convention of CDMA and call the basic elements in
( ) K
d and
( ) K
x chips. Spreader block can be either the same or
different for different users [11]. From a performance point of
view, it is advantageous to use a low-rate FEC code before
spreader block [7] that can provide an extra coding gain.

The key principle of IDMA is that the interleavers {
( ) K
+ }
should be different for individual users. We assume that the
interleavers are generated independently and randomly. For
simplicity, we first consider time-invariant single-path channels
with real channel coefficients and BPSK signaling.
After sampling at chip rate, the received signal from K users can
be written as
( ) ( )
1
m
K K
j j
r h x =
j
n + (1)
j = 1, 2, , J,
where
( ) K
j
x is the jth chip transmitted by the Kth user, the
channel coefficient for the Kth user and { } samples of a zero-
mean additive white Gaussian noise (AWGN) with variance
j
n
2
o =N
0
/2.

In the lower part of Fig. 1, which shows the receiver section of
IDMA system, the iterative multiuser detector employs a low-cost
chip-by-chip detection strategy and avoids conventional MAP. The
receiver consists of Elementary Signal Estimator (ESE) and an a
posteriori probability (APP) decoder (DEC). The ESE exchanges
information with the DEC in a turbo-type manner [8]. The DEC
produces hard decisions {
K
rx
d } on information bits {
K
d } in the
final iteration.

The receiver operation is based on the received signal r { , j =
1,, J}, and the channel coefficients are denoted by h[
,, ,.,
j
r
(1)
h
( ) k
h
( ) K
h ]. Finding an optimal solution is usually
prohibitively complicated. We now consider a sub-optimal
approach by first separating the conditions, i.e., r, and h, and then
combining the results using an iterative process. This greatly
reduces the complexity involved.

Specifically, the constraint of Coder is ignored in the ESE. The
output of the ESE is defined by the logarithm likelihood ratio
(LLR)
( | ( ) 1, )
{ ( (
K
j))}
( | ( ) 1, )
j K
j K
p r x j h
e x
p r x j h
log
ESE
( =+
=
(
=
(

(2)
, K j

The DEC consists of K local APP decoders. The kth local APP
decoder performs an APP decoding of signal for the kth user
using , after appropriate deinterleaving, as its input. Its output
is the so-called extrinsic LLR [8].
ESE
e
( (
( (
r K
r K
p x
p x
) 1| ( ( )))
log ( ( ))
) 1| ( ( )))
ECE K
K ECE K
ECE K
j e x j
e x e x j
j e x j
( =+
=
(
=

{ (
DEC
, K
( ))} j
j

(3)
The APP decoding in the DEC is a standard function [8] so we will
not discuss it in detail. In the following, we will focus on the ESE.

Descriptions of the ESE

The ESE generates coarse estimates of { ( )
K
x j , j = 1,, J, K =
1,..k,.., K}. We ignore the constraint of Coder to maintain low
complexity. Consider the jth chip of the kth user
with ( )
K
x j e{+1, -1}under BPSK modulation. We treat
(
K
) x j as a random variable and use (initialized
to zero) to approximate the a priori LLR of
( (
DEC K
e x
(
K
)) j
) x j ,
( ( ) 1
{ ( ( ))} log
( ( ) 1
r K
r K
p x j
j
p x j
DEC K
e x
( = +
~
(
=

(4)
Based on (4), we have
exp(
exp(
(
K
e x
( ( ))) 1
( (
( ( ))) 1
DEC K
K
DEC K
e x j
e x j

=
+
))) j
nh((
DEC
E x
ta
(5)
( ))) / 2) j = (6)
and
2
Va ) ( ( )))
K
r( ( ))=1-
K
x j E x j , (7)



III. CORRELATION OF INTERLEAVERS

An obvious interleaver design criterion is that every two
interleavers out of a set of interleavers should not collide since
the separation of users is achieved by interleavers [9].

844
International Conference on Information and Communication Technology (IICT-07), July 26-28, 2007
Unlike in classical turbo coding/decoding, where the task of a
single interleaver is to decorrelate different sequences of bits, here
we have a set of interleavers that not only need to decorrelate
different bit sequences, but also different users. By correlation
between interleavers, it is meant that how strongly signals from
other users affect the decoding process of a specific user. Hence,
the additive noise should not play a role in the correlation of
interleavers, and hence, IDMA system, may be considered to be
noise-less.

In mathematics, orthogonal is synonymous with perpendicular i.e.
inner product of two vectors is zero. From [9], the orthogonality
can be defined as no correlation between two interleavers out of set
of interleavers. So, in IDMA scheme, in order to distinguish the
users, every user must have interleavers with no correlation. In
other words, the interleavers must be orthogonal to each other in
order to have no collision.

Let and
i
+
j
+
be two interleavers and let w and v be two words.
Now, we may define the correlation C( , w,
i
+
j
+
, v) between
i
+

and
j
+
with respect to the words w and v as the scalar product
between
+
(f(w)) and
i j
+
(f(v)):
C( , w,
i
+
j
+
, v) ={ (f(w)),
i
+
j
+
(f(v))}=0 (8)


IV. INTERLEAVER DESIGN


The proposed interleaver is based on the combination of cyclic and
diagonal interleavers. We will first see the cyclic and diagonal
interleavers in brief. Later the basic concept of proposed
interleaver will be explained.
In cyclic interleaver, the elements are arranged in a sequence with
increasing from 1 to N. Total number of elements N is divided into
two parts N
1
and N
2
such that N= N
1.
N
2
, where N
1
and N
2
are
referred as rows and columns respectively.
Let us take N=15 while N
1
=5 and N
2
=3. So, the arrangement of the
sequence will be as shown below in the form of matrix.

1 4 7 10 13
2 5 8 11 14
3 6 9 12 15

)
(9)

For cyclic interleaver, a delay element is introduced in the rows of
the matrix. The delay is chosen to be less than half of N
1
. In our
example, for N
1
=5, the delay is chosen to be 2. So the rearranged
version of matrix will be,

1 4 7 10 13
11 14 2 5 8
6 9 12 15 3

)
(10)

Reading the matrix column-wise, we get the data as following


1 11 6 4 14 9 7 2 12 10 5 15 13
8 3


For diagonal interleaver, we first take the matrix given in (9), and
read the data in a sequence as shown in (11).
(11)

Now, the data is read as shown by lines. Hence, the data will be
read as

1 5 9 4 8 12 7 11 15 10 14 3 13 2
6


The data sequences found after both type of interleaving does not
support orthogonality for more than one user but if the mechanism
of both of the interleaver is combined, we get new type of
interleaver called as proposed interleaver which supports
randomness and orthogonality in limited number user area with
user not exceeding 20 at a time.
The data in proposed interleaver is read in diagonal manner as it
was in case of diagonal interleaver in (11). So, the data may in the
form as shown below.
1 14 12 4 2 15 7 5 3 10 8 6 13 11
9
At the receiver, the data have to be turbo processed before
decoding. The working of the receiver shown in fig. 1 is evident in
section II.
In the next section, we will present the simulation results for fixed
number of iterations, data length and spread length. The simulation
is done with matlab environment.

V. SIMULATION RESULTS

It is well known that the performance of CDMA goes down with
increment in count of users operating at a particular time in an
area. Even its performance is not better at 16 users in comparison
that of IDMA [11]. The IDMA scheme performs better in terms of
number of user in comparison to CDMA [11] where random
interleaver [2] and master random interleaver [14].
The master interleaver used in [14] was designed to alleviate the
requirement of higher memory which was needed to store the
information about interleavers for individual users.


845
International Conference on Information and Communication Technology (IICT-07), July 26-28, 2007
3 4 5 6 7 8 9
10
-5
10
-4
10
-3
10
-2
10
-1
10
0
Eb/No(dB)
B
i
t

E
r
r
o
r

R
a
t
e


n=4
n=12
n=20

Figure 2. Uncoded IDMA with BPSK modulation with data
length=1024, spread length=16, and iteration=20 with different
count of users

The simulation results shows that proposed interleaver performs
better than conventional CDMA and existing IDMA simulations in
terms of lower BER for same number of users.

The proposed interleaver performs better at number of users not
exceeding 20. For higher count of users, the performance gets
degraded. So, this interleaver is recommended to be used with
IDMA scheme with are where less than 20 users are operative at a
time.

If we compare the simulation results for uncoded IDMA scheme
with BPSK modulation with different number of users, it is
observed that with increment in user count, the performance of the
IDMA scheme suffers. It shows that randomness of the proposed
interleaver is getting lessened for higher count of the users. The
computer simulation is shown in figure 2. The other computer
simulation shown in figure 3 includes data length to be 1024 while
spread length is 16 with 20 iterations.

The results are shown in figure 3 for 20 users operating at a time
for different types of interleavers including random interleaver [2],
master random interleaver [14], and the proposed interleaver. It is
evident from the simulation that the performance of IDMA scheme
gets improved with the proposed interleaver.

The results of the simulation shows that the proposed interleaver is
best suited for the areas where the count of the user is less than 20.
Also
1 2 3 4 5 6 7 8 9 10
10
-6
10
-5
10
-4
10
-3
10
-2
10
-1
Eb/No(dB)
B
i
t

E
r
r
o
r

R
a
t
e


randominterleaver
master randominterleaver
proposed interleaver

Figure 3. Uncoded IDMA with BPSK modulation with data
length=1024, spread length=16, and iteration=20 for number of
users =20

it performs much better than existing CDMA system under similar
conditions [11].

VI. CONCLUSION

In [1], it is shown that IDMA with random interleavers can support
more users than a comparable CDMA system. The main
contribution of this work is to design a new interleaver which is
having similar performance in terms of BER also having lower
memory requirement as well easier to generate and implement. The
simulations in Section V show that the performance of proposed
interleaver is very similar to the performance of random interleaver
in [2, 14]. However, the interleaver assignment scheme is
simplified and memory cost is greatly reduced. Lesser bits are
needed to be sent over channel for transmission in comparison to
interleaver presented in [14].

Since the performance of the interleaver described in Section IV is
very similar to the performance of random interleaver, we can
affirm that IDMA using proposed interleaver, can support better
than both conventional CDMA system and IDMA system with
random interleaver, for limited user area network.


REFERENCES

1) Li, P., et al.: Approaching the capacity of multiple
access channels using interleaved low-rate codes, IEEE
Commun. Lett., 2004, 8, (1), pp. 46
2) Li, P., et al.: Interleave-division multiple-access, IEEE
Trans. Wireless Commun., vol. 5, no. 4, pp. 938-947,
April 2006, pp. 938-947.
3) Moher, M.: An iterative multiuser decoder for near-
capacity communications, IEEE Trans. Commun., 1998,
46, (7), pp. 870880
4) Wang, X., and Poor, H.V.: Iterative (turbo) soft
interference cancellation and decoding for coded
846
International Conference on Information and Communication Technology (IICT-07), July 26-28, 2007
847
CDMA, IEEE Trans. Commun., 1999, 47, (7), pp.
10461061
5) L. Ping, L. Liu, W. K. Leung, A Simple Approach to
Near-Optimal Multiuser Detection: Interleave-Division
Multiple-Access, WCNC 2003 - IEEE Wireless
Communications and Networking Conference, vol. 4,
no.1, Mar 2003, pp. 391-396.
6) M. Shukla, V.K. Srivastava, S. Tiwari, Interleave
Division Multiple Access : New trend in Wireless
Communication, in Proc. NCCT- Dehradun, India-2006,
pp. 151-154.
7) Li Ping, L. Liu, K. Y. Wu, and W. K. Leung, A unified
approach to multi-user detection and space time coding
with low complexity and nearly optimal performance, in
Proc. 40th Allerton Conference, Allerton House, USA,
Oct. 2002, pp. 170179.
8) C. Berrou and A. Glavieux, Near Shannon limit error
correcting coding and decoding: Turbo-codes, IEEE
Trans. Commun., vol. 44, pp. 12611271, Oct. 1996.
9) Ioachim Pupeza, Aleksandar Kav and Li Ping ,Efficient
Generation of interleavers for IDMA, Proc. IEEE Int.
Conf. on Commun., ICC'06, Istanbul, Turkey, 11-15
June 2006.
10) K. Y. Wu, W. K. Leung and Li Ping, A simple approach
to near-optimal multiple transmit antenna space-time
codes, IEEE Intern. Conf. on Commun., ICC'03, pp.
2603-2607.
11) Li Ping, Lihai, Liu and W. K. Leung, A simple
approach to near-optimal multiuser detection: interleave-
division multiple access, IEEE Wireless
Communications and Networking Conference,
WCNC'03, pp. 391-396.
12) Li Ping, Lihai Liu, K.Y. Wu and W. K. Leung, A
unified approach to multiuser detection and space-time
coding with low complexity and nearly optimal
performance, 40th Allerton Conference, Allerton House,
USA, Oct 2002, pp. 170-179.
13) Li Ping, Interleave-division multiple access and chip-
by-chip iterative multi-user detection, IEEE Commun.
Magazine, vol. 43, no. 6, pp. S19-S23, June 2005.
14) H. Wu, Li Ping and A. Perotti, User-specific chip-level
interleaver design for IDMA systems, IEEE Electronic
Letters, vol. 42, no. 4, Feb 2006.
15) H. Schoeneich and P. Hoeher, Adaptive interleave-
division multiple access A potential air interface for
4G bearer services and wireless LANs, in Proc. WOCN
2004, Muskat, Oman, June 2004, pp. 179182.
16) L. Liu, W. K. Leung, and Li Ping, Simple chip-by-chip
multi-user detection for CDMA systems, in Proc. IEEE
VTC-Spring, Korea, Apr. 2003, pp. 2157-2161.

Potrebbero piacerti anche