BLIND DESPREADING OF WCDMA SIGNALS
ZHAO Zhi-jin
a,b
, GU Xiao-wei
a
, SHEN Lei
a
, ZHAN Yi
b
Zhejiang Provincial Key Lab of Data Storage and Transmission Technology, Hangzhou Dianzi
University, Hangzhou Zhejiang 310018 a
State Key Lab of Information Control Technology in Communication System of No.36 Research
Institute,China Electronic Technology Corporation, Jiaxing Zhejiang 314001 b
zhaozj03@hdu.edu.cn, johnsongutc@126.com, shenlei@ hdu.edu.cn
ABSTRACT
Aiming at despreading Wideband Code Division Multiple
Access (WCDMA) signal blindly, the phase of Gold
sequence is estimated firstly by utilizing the maximum value
of the fourth-order cumulant slice of the product of
WCDMA signal with the optimum m-sequences of different
delays, and then the signal is descrambled. With the
utilization of particular structure of OVSF codes, the signal
is synchronized through the maximum value of cross-
correlation function of signal’s segments. By calculating the
cross-correlation of each row of Hadamard matrix with the
signal that has been descrambled and synchronized before,
the OVSF code is obtained from the column of the matrix
corresponding to the largest value of the cross-correlation
function, and the signal is despread finally. Simulation
results show the better performance of the proposed method
in despreading WCDMA signals under the background of
additive white Gaussian noise.
Index Terms—Wideband Code Division Multiple Access
(WCDMA) signals, Fourth-order cumulant slice, Blind
despreading, Orthogonal Variable Spreading Factor (OVSF)
code estimation
1. INTRODUCTION
Due to its exceptional resistance to inference and strong
confidentiality, Wideband Code Division Multiple Access
(WCDMA) communication systems are widely used in civil
and military applications. The need for blind despreading
the WCDMA signals in non-cooperative communications
becomes increasingly urgent.
References [1,2] investigate the blind despreading
methods of periodic long code direct sequence spread
spectrum (DS-SS) signals. Through segmentation
processing and average cross-correlation calculation, the
starting symbol of PN sequence is determined, and PN
sequence is estimated by utilizing the eigen-analysis
technique in reference [1]. In reference [2], the synchronous
multi-user long-code DS-SS signals are converted as the
short-code ones with missing data, and then the users’
spreading waveform subspace is estimated by utilizing the
Singular Value Thresholding (SVT) algorithm, and finally
the Expectation Maximization (EM) algorithm is used to
blindly despreading the signals.
Despite of the strong need, few methods of detection
and blind despreading of WCDMA signals are published at
present, and most of which use higher order statistics to
detect signal because of its inhibition effect on additive
white Gaussian noise (AWGN)
[3],[4]
. WCDMA signals use
Orthogonal Variable Spreading Factor (OVSF) codes as
spreading codes, which have ideal orthogonal property only
when signals are synchronous
[5]
. Reference [6] mainly uses
eigenvalue decomposition (EVD) to estimate OVSF codes
of WCDMA signals. As WCDMA signals are scrambled by
Gold sequences, EVD method can not be applied to signals
that contain two pseudo-random (PN) codes, a method about
blind despreading WCDMA signal has not been seen yet.
In this paper, an investigation of blind descrambling
and despreading method for WCDMA signals is conducted
when primitive polynomials of optimum m-sequences pair
of Gold sequence and period of OVSF code are known. The
rest of the paper is organized as follows. In section 2, the
signal model of WCDMA signal is given. The blind
despreading algorithm of WCDMA signal is proposed in
section3. Simulations and analyses are presented in section
4 and finally, the concluding remarks are given in section 5.
2. SIGNAL MODEL
The WCDMA signal is supposed to use the values +1, -1
instead of 0 and 1. After sampling the signal with chip rate
of OVSF code, the received base-band WCDMA signal can
be expressed as
()rn , 0,1, 2, 1nN=−" :
() ()()() ()rn Adnongn wn
=+ (1)
where,
A is the amplitude of the signal, n denotes the
sampling time. ( )
dn is the data sequence in signal, each
symbol is sampled by
O
times. Walsh code generated by
Hadamard matrix with dimension of
O
is used as the
978-1-4799-2186-7/14/$31.00 ©2014 IEEE