Differential Transmit Diversity Based on
Quasi-Orthogonal Space-Time Block Code
Chau Yuen, Yong Liang Guan
School of Electrical and Electronic Engineering
Nanyang Technological University
Singapore
yuenchau@pmail.ntu.edu.sg, eylguan@ntu.edu.sg
Tjeng Thiang Tjhung
Institute for Infocomm Research
Singapore
tjhungtt@i2r.a-star.edu.sg
Abstract— By using joint modulation and customized
constellation set, we show that Quasi-Orthogonal Space-Time
Block Code (QO-STBC) can be used to form a new differential
space-time modulation (DSTM) scheme to provide full transmit
diversity with non-coherent detection. Our new scheme can
provide higher code rate than existing DSTM schemes based on
Orthogonal STBC. It also has a lower decoding complexity than
the other DSTM schemes, such as those based on Group Codes,
because it only requires a joint detection of two complex symbols.
We derive the design criteria for the customized constellation set
and use them to construct a constellation set that provides a wide
range of spectral efficiency with full diversity and maximum
coding gain.
Keywords - differential space-time modulation; low decoding
complexity; non coherent detection; quasi-orthogonal space-time
block code.
I. INTRODUCTION
In wireless communications, system performance is often
severely degraded by fading effects due to multi-path signal
propagation. Modulation techniques designed for multiple
transmit antennas, called space-time modulation or transmit
diversity can be used to reduce fading effects effectively. Early
transmit diversity schemes were designed for coherent
detection, with channel estimates assumed available at the
receiver. However, the complexity and cost of channel
estimation grow with the number of transmit and receive
antennas. Therefore the availability of transmit diversity
schemes that do not require channel estimation is desirable. To
this end, several differential space-time modulation (DSTM)
schemes have been designed [1-8].
Hughes has designed a DSTM based on group codes [1, 2];
Hochwald and Sweldens have designed a DSTM based on
unitary matrices [3]. Tarokh and Jafarkhani have proposed a
DSTM by using the Orthogonal Space-Time Block Code (O-
STBC) [4], while Ganesa and Stoica [8] provide another
DSTM also based on O-STBC but with a much simpler
decoding complexity. Hassibi et al. has proposed DSTM based
on Cayley code and Sp(2). Recently, Al-Dhahir gives a new
rate-two DSTM based on two parallel Alamouti O-STBCs for
four transmit antennas system [6]. Among them the scheme in
[8] has the simplest decoding complexity. However the DSTM
in [8] is based on the O-STBC, whose maximum achievable
code rate is limited to ¾ for four antennas and ½ for eight
antennas when used with complex constellations.
In this paper, we propose a new DSTM scheme that is
based on Quasi-Orthogonal STBC (QO-STBC) [9-13], to
provide full transmit diversity with higher code rate than that of
[8] and lower decoding complexity than other DSTM schemes
[1-7]. We will derive the design criteria and construct an
example constellation set for the proposed DSTM scheme. We
will also study its decoding complexity and decoding
performance in comparison with the existing schemes.
The organization of this paper is as follow, Section II
reviews the DSTM signal model and the decoding performance
criteria of DSTM. Section III proposes the new DSTM scheme
based on QO-STBC. And Section IV gives the performance
comparison. Finally Section V concludes the paper.
II. R
EVIEW OF DSTM
A. DSTM Signal Model
Consider a MIMO communication system, with N
T
transmit antennas and N
R
receive antennas. Let H
t
be the N
R
×
N
T
channel gain matrix at a time t. Thus the ij
th
element of H
t
is
the channel coefficient for the signal path from the j
th
transmit
antenna to the i
th
receive antenna. Let C
t
be the N
T
× P
codeword transmitted at a time t. Then, the received N
R
× P
signal matrix R
t
can be written as
tttt
=+RHCN
(1)
where
N
t
is the additive white Gaussian noise. In this paper, the
code length P is set equal to N
T
as in [1] and [8], so that the
transmitted codeword
C
t
is a square matrix.
At the start of the transmission, we transmit a known
codeword C
0
, which is a unitary matrix of size N
T
× N
T
. The
codeword
C
t
transmitted at a time t is differentially encoded by
1ttt−
=CCU
(2)
where U
t
is a unitary matrix (such that U
t
U
t
H
=I) called the code
matrix, that contains information of the transmitted data. Since
Globecom 2004 545 0-7803-8794-5/04/$20.00 © 2004 IEEE
IEEE Communications Society