Robust Waveform Design for MIMO-STAP with
Imperfect Clutter Prior Knowledge
Hongyan Wang
1
, Bingnan Pei
2
, Yunfeng Bai
3
College of Information Engineering
,
Dalian University
Dalian, China
1
gglongs@163.com,
2
peibn56@sohu.com,
3
baiyunfeng@dlu.edu.cn
Abstract—In this paper, we address the problem of robust
waveform optimization with imperfect clutter prior knowledge to
improve the worst-case detection performance of multi-input
multi-output (MIMO) space-time adaptive processing (STAP) in
the presence of colored Gaussian disturbance. An iterative
algorithm is proposed to optimize the waveform covariance
matrix (WCM) for maximizing the worst-case output signal-
interference-noise-ratio (SINR) over the convex uncertainty set
such that the worst-case detection performance of MIMO-STAP
can be maximized. By exploiting the diagonal loading (DL)
method, each iteration step in the proposed algorithm can be
reformulated as a semidefinite programming (SDP) problem,
which can be solved very efficiently. Numerical examples show
that the worst-case output SINR of MIMO-STAP can be
improved considerably by the proposed method compared to
that of uncorrelated waveforms.
Index Terms—MMO-STAP, robust waveform design,
diagonal loading, semidefinite programming.
I. INTRODUCTION
In [1], H. Wang et al. has investigated the problem of
waveform design for multi-input multi-output (MIMO) space-
time adaptive processing (STAP) to maximize the output
signal-interference-noise ratio (SINR) to improve the
detection performance. It is obvious that the optimization
problem in this document requires the specification of
parameters, e.g., the target location, the clutter channel, etc..
As a sequence, the optimized waveforms depend on these pre-
assigned values. In practice, these parameters are estimated
with errors, and hence they are uncertain. As illustrated by
numerical examples in [1], the resultant output SINR, i.e.,
detection probability, is sensitive to these estimation errors
and uncertainty in parameters, which is similar to that in the
case of waveform design for improving the parameter
estimation [2]. It means that the optimized waveforms based
on a certain parameter estimate can give a very low detection
performance for another reasonable estimate.
In order to improve the worst-case detection performance
in the presence of colored Gaussian disturbance, the problem
of robust waveform design with imperfect clutter prior
knowledge is addressed in this paper, which attempts to
systematically alleviate the sensitivity by explicitly
incorporating a parameter uncertainty model in the
optimization issue. Because maximization of the output SINR
is tantamount to maximization of the detection performance in
the case of Gaussian noise [3], here the waveform covariance
matrix (WCM) is optimized to maximize the worst-case
output SINR of MIMO-STAP over the convex uncertainty set
such that the worst-case detection performance can be
maximized. An iterative algorithm is proposed to solve the
optimization problem. By using the diagonal loading (DL)
method [4], each step in the proposed algorithm can be
reformulated as a semidefinite programming (SDP) problem
[5], and hence it can be solved efficiently.
The rest of this paper is organized as follows. The MIMO-
STAP model is introduced, and the robust optimization
problem is formulated in SectionⅡ. An DL based iteration
algorithm is proposed to formulate the resultant nonlinear
optimization problem as an SDP in Section Ⅲ . The
effectiveness of the proposed method is verified via numerical
examples in Section Ⅳ . Finally, conclusions are given in
SectionⅤ.
II. P
ROBLEM FORMULATION
The MIMO-STAP signal model adopted in this paper is the
same as that in [1]. For the lth pulse repetition interval (PRI),
the data received by all the receiving elements can be
expressed as:
,
1
2
2
0
πβ
π
ρρ
−
=
=+ +
∑
YabS abSZ
C
si
D
N
jfl
jfl T T
lt i ii l
i
ee, (1)
where a and
a
i
denote, respectively, the receiving steering
vectors for the target and the clutter patch at
θ
i
; b and
b
i
denote, respectively, the transmitting steering vectors for
the target and the clutter patch at
i
θ
; and Z
l
denotes the
interference plus noise received by all the receiving elements
in the
lth PRI. Similar to [2], we make a simplifying
assumption that the columns of
Z
l
are independent and
identically distributed circularly symmetric complex Gaussian
random vectors with mean zero and an unknown covariance
Q .
The sufficient statistics for STAP signal processing can be
extracted by a bank of matched filters
1
2
()
−
SSS
HH
, which
can be stacked in a 1×MN vector as:
,
1
2
2
1
1
2
2
0
()()
()()vec()
π
πβ
ρ
ρ
−
=
=⊗⊗+
⊗⊗+
∑
S
S
yRIba
RIba Z
D
C
si
T
jfl
lt N
N
T
jfl
iNiil
i
e
e
, (2)
978-1-4799-5274-8/14/$31.00 ©2014 IEEE