Space-Time Block Code Design for
Asymmetric-OFDM Systems
Lin Luo
Institute for Telecommunication Research, University of South Australia, Australia
Email: Lin.Luo@unisa.edu.au
Abstract—A considerable drawback of OFDM systems
is their high peak-to-average power ratio (PAPR). The
Asymmetric-OFDM (A-OFDM) system addresses this problem
by reducing the size of fast Fourier transform (FFT) / inverse
FFT (IFFT). Based on the layered concept in FFT algorithms
and Alamouti-like space time block code (STBC), we extend
the A-OFDM to an full diversity MIMO multicarrier system
and apply it to vehicle-to-vehicle communications. Comparing
with the conventional OFDM with STBC, the proposed system
can be arbitrarily adjusted in the number of subchannels
(or the length of each subchannel), typically in order to
suit requirements for high reliability or low complexity, in
different applications and scenarios. Theoretical analysis in
a frequency selective channel and simulations in a vehicle-
to-vehicle environment show that the proposed system, with
any adjustable size of subchannels from 1 to N , N being
the transmission block length, has higher reliability than an
equivalent MIMO-OFDM system.
I. INTRODUCTION
The asymmetric OFDM system (A-OFDM) [1], [2] is
a low-complexity trade-off solution between OFDM and
SC-FDE systems in terms of peak-to-average power ratio
(PAPR) and carrier frequency offset (CFO) sensitivity, as
well as bit error rate (BER) performance. The A-OFDM
can be freely adjusted in the length of subchannels, P ,
in a broad range between OFDM (P =1) and SC-FDE
(P = N, N being the transmission block length), according
to requirements of PAPR, CFO sensitivity, BER performance
and complexity. In [2] it has been shown that the PAPR
reduces for higher P . On the other hand, the complexity
decreases with lower P [2].
A reliable transmission is required in safety applications
of vehicle-to-vehicle communications. In order to improve
the reliability, i.e. reduce the BER, one option is to use
multiple antennas at the transmitter and receiver in order to
benefit from space diversity. Multiple-input multiple-output
(MIMO) systems for OFDM have already been well studied
[3], [4], [5]. A first proposition of a MIMO version of
the Quadrature-OFDMA (Q-OFDMA) system, which is a
multiple access version of A-OFDM, has been given in
[6], using an Alamouti-like space time block code (STBC).
Based on [6], this paper elaborates a general multi-carrier
MIMO system with adjustable length of subchannel, variable
PAPR and variable complexity of transmitter and receiver,
and analyze its BER performance based on zero forcing (ZF)
and minimum mean square error (MMSE) equalizations. The
proposed system has reduced PAPR and outperforms the
OFDM system in terms of BER performance. This makes it
an interesting candidate for applications in dedicated short
range communications (DSRC).
The rest of this paper is organized as follows. In Section II
we further develop the combining scheme at the receiver for
Alamouti-like STBC to achieve maximal diversity, and give
a general system for a 2 × n
r
STBC using ZF and MMSE
equalizations, where n
r
is the number of receive antennas. In
Section III, the BER performance is theoretically analyzed
in a Rayleigh fading channel for uncoded MIMO A-OFDM
systems. In Section IV this new MIMO A-OFDM system
will be simulated for vehicle-to-vehicle communications on
a geometry-based stochastic channel model. We compare it
to a single antenna system as well as to conventional MIMO-
OFDM systems. Finally, Section V concludes this paper.
II. S
PACE-TIME BLOCK CODING FOR A-OFDM
A. System model for A-OFDM
Given two N-point time-domain symbols x, h, and their
noiseless circular convolution output y = x h, their
frequency-domain correspondences via DFT have the rela-
tionship
˜
y =
˜
x
˜
h. If we rearrange the frequency domain
symbols
˜
x,
˜
h and
˜
y into P × Q matrices (PQ = N)row-
wise according to the layered IFFT structure concept [2],
the vectors
˜
x
q
,
˜
h
q
and
˜
y
q
from the q-th column of the
matrices retain that
˜
y
q
=
˜
x
q
˜
h
q
, where [
˜
y
q
]
p
=[
˜
y]
pQ+q
,
[
˜
x
q
]
p
=[
˜
x]
pQ+q
, [
˜
h
q
]
p
=[
˜
h]
pQ+q
, and p =0, 1, ···,P − 1.
We now introduce the intermediate-domain symbols
{
˘
x
q
,
˘
h
q
,
˘
y
q
} as the IDFTs of {
˜
x
q
,
˜
h
q
,
˜
y
q
}, given by
˘
x
q
=
F
H
P
˜
x
q
,
˘
h
q
= F
H
P
˜
h
q
,
˘
y
q
= F
H
P
˜
y
q
, where F
H
P
is the normalized
P -point IDFT matrix. According to the convolution property
of DFT, we have
˘
y
q
=
˘
x
q
˘
h
q
, which establishes the
relationship of symbols in the intermediate domain, and can
be expressed in matrix form as
˘
y
q
=
˘
H
q
˘
x
q
, where the P ×P
circulant matrix
˘
H
q
represents the dispersive channel, with
[
˘
H
q
]
i,j
=
˘
h (((i − j) modP ) Q + q), where
˘
h(·) denotes the
channel response in the intermediate domain.
At the receiver of the A-OFDM system, in order to realize
a one-tap equalization, the weighting outputs are transformed
from the intermediate domain to frequency domain as
˜
y
q
= F
P
˘
y
q
= F
P
˘
H
q
˘
x
q
+ F
P
˘
n
q
= D
q
F
P
˘
x
q
+
˜
n
q
, (1)
where we now include
˘
n
q
and
˜
n
q
, the white Gaussian noise
in intermediate and frequency domain respectively, and D
q
is
the diagonal matrix with the channel frequency coefficients
˜
h
p,q
on its diagonal. In the above formulas we consider a
GC'12 Workshop: The 8th Broadband Wireless Access Workshop
978-1-4673-4941-3/12/$31.00 ©2012 IEEE 204