Published in IET Signal Processing
Received on 21st February 2014
Revised on 27th May 2014
Accepted on 29th July 2014
doi: 10.1049/iet-spr.2014.0079
ISSN 1751-9675
Frequency offset estimation of the linear mixture
of two co-frequency 8 phase-shift keying
modulated signals
Yong Yang, Dongling Zhang, Hua Peng
China National Digital Switching System Engineering and Technological Research Center, Zhengzhou, Henan,
People’s Republic of China
E-mail: yangyong328@163.com
Abstract: A frequency offset estimation (FOE) algorithm based on the maximum-likelihood criterion is proposed for the linear
mixture of two 8PSK modulated signals with fixed frame length and frame synchronisation code on the condition that the frame
length is known when the signals are coded asynchronously. FOE is calculated by optimising the objective function, which is
obtained from cross-correlation computation, given that timing error, amplitude attenuation and phase offset are unknown.
The optimised process is accomplished by utilising hierarchical search. Cramer-Rao bound (CRB) of mixed-signal FOE is
derived to evaluate the performance of the proposed algorithm. The estimation performance of the algorithm can come close
to the CRB performance with low E
s
/N
0
. The proposed algorithm is also applied to the FOE of a mixed signal whose power
is asymmetrical.
1 Introduction
Paired carrier multiple access (PCMA) is a new satellite
communication technolog y [1]. The signals sent by two
earth stations are overlapped both in time and frequency
domains. Hence, PCMA can save half of the bandwidth
resources and improve the performance of anti-interception.
The demodulation of PCMA signals is fundamentally a
field in single-channel blind source separation [2–5]or
single antenna interference cancellation [6–8]. Therefore
PCMA signals have become a challenge in communication
signal processing.
As a special characteristic of PCMA technology, carrier
synchronisation is the most significant link in demodulation,
with significant achievements obtained in single digital
modulation signals. The most common method of
demodulation is performed by eliminating the modulated
information through M-power operation and estimating the
frequency offset based on the maximum-likelihood (ML)
criterion [9–11]. This method can also be applied to
low-order modulated PCMA signals, such as those in
binary phase shift keying and quadrature phase shift keying.
However, the presence of too many cross terms after
M-power operation for 8 phase-shift keying (8PSK)
modulated mixed signals renders the ML estimation
algorithm inapplicable. Thus, demodulating the mixed
signal of two 8PSK modulated signals has become a
challenge in signal processing. The frequency offset
estimation (FOE) of two 8PSK modulated signals requires
an immediate solution.
An FOE algorithm based on ML for sync codes is proposed
in this study. The algorithm is applied when two 8PSK
modulated signals are coded asynchronously. The proposed
algorithm has low computational complexity given that
timing error, amplitude attenuation and phase offset
information are unknown. Thus, the algorithm has a good
potential for future application.
This paper is organised as follows. The basic signal model
for the linear mixture of two co-frequency 8PSK modulated
signals is described in Section 2. The FOE algorithm based
on ML is proposed in Section 3. Section 4 focuses on the
derivation of the CRB of the mixed signal. Section 5
provides the simulation results, which show the
effectiveness of the proposed algorithm. Lastly, conclusions
are provided in Section 6.
2 Signal model
The received linear mixture of two co-frequency 8PSK
modulated signals that have the same symbol rate is
considered in satellite communication. The received signal
is sampled at symbol period T, and the discrete-time form
can be written as
y
k
=
2
i=1
h
i
e
j(2
p
f
i
kT+
w
i
)
x
(i)
k
+ v
k
(1)
where v
k
(k=1, 2, …) is the complex additive white Gaussian
noise (AWGN) sampled sequences with zero mean and
www.ietdl.org
186
&
The Institution of Engineering and Technology 2015
IET Signal Process., 2015, Vol. 9, Iss. 2, pp. 186–192
doi: 10.1049/iet-spr.2014.0079