The Research of MUSIC-AML Algorithm for MIMO
Radar System
Li Shufeng Zhang Yuchi Zhang Yunfeng
School of Information Engineering, Communication University of China, Beijing, 100024, China
lishufeng@cuc.edu.cn
Abstract—Direction-of-arrival (DOA) is a research direction in
array signal processing. This paper estimates the DOA for
Multiple-Input Multiple-Output (MIMO) radar system. The
algorithms of multiple signal classification and approximated
maximum likelihood (MUSIC-AML) based on linear array are
deduced. The number of target sources is also valued in this
paper. The results show the feasibility and validity of the
algorithm applied in MIMO radar system.
Keywords: MUSIC; DOA; MIMO
I. INTRODUCTION
In radar domain, the most normal array based on
direction-of-arrival (DOA) estimation algorithm is
conventional beam forming algorithm, however, this method
is restricted by antenna aperture length. The array aperture
cannot be manufactured so large enough. The multiple signal
classification (MUSIC) algorithm and
approximated
maximum likelihood (AML)
can be applied to solve above
problem[1-2]. The orthogonal characteristic between signal
sub-space and noise sub-space is applied in the array signal
processing. Spectral peak with nail shape is constructed
significantly to increase the target resolution of algorithm in
the condition that the number of channel sources is predicted.
The available array estimation algorithms are only presented
for receiving array, while in MIMO radar, receiving beam
forming in addition to transmitter beam forming, equaling to
increase the antenna aperture when the receiving beam is
formed, thus the angular measurement resolution is
increased.
Many works are done in DOA algorithm under different
clutter background in MIMO Radar. The signal and clutter
model for MIMO radar in addition to average Cramer-Rao
bound (CRB) under composite gauss clutter background are
given in [3-4]. The DOA estimation algorithm based on
Fractional lower order minimum variance distortion less
response and the infinite norm normalized minimum
variance distortion less response algorithm is given in[5].
According to different angular tracking condition, the
maximum likelihood estimation method is presented in[6-9].
The data are processed from both end, and searched in only
one dimension. Based on the above research work, the
MUSIC-AML algorithm is derived and analyzed referring to
linear array in MIMO radar in this paper.
II. M
IMO ANTENNA ARRAY MODEL
Assuming that the transmitting and receiving arrays are
linear arrays and located in the type in Fig.1, the transmitting
array is composed of M elements, and the distance between
adjacent elements is dt, while the receiving array is
composed of N elements, and the distance between adjacent
elements is dr. Assuming that the signal is located in far field,
and the electric wave from the signal to each array element
is plane wave.
Fig.1. MIMO Antenna Array Model
In practical condition, the space angle of arrival wave
should be expressed in three dimension space [10]. While in
order to illustrate intuitively, the vector of arrival wave is
restricted within the plane and the angle between plane wave
and normal of linear array is DOA. If the transmitting and
receiving arrays are separate, the angle between target and
array is different. For further simplification, we assume that
the relative position of transmitting and receiving array is
fixed in advance, so the direction from target to transmitting
and receiving array is identical.
The transmitting direction vector of average array is
described as follows:
22
sin ( 1) sin
a( ) 1, , ,
rr
T
dd
jjN
ee
(1)
While the receiving vector is
22
sin ( 1) sin
b( ) 1, , ,
tt
T
dd
jjM
ee
(2)
The steering vector of MIMO radar is
() () ()
Aab (3)
III. M
USIC-AML ALGORITHM
If the number of array elements is greater than the
number of signal sources, the signal of the array matrix data
forms a subspace, which is called signal subspace;
uncorrelated with the signal. The noise subspace composed
by the irrelevant signal is orthogonal with the signal
subspace. The orthogonality may determine the signal DOA.
MUSIC is a signal parameter estimation algorithm that has
high resolution. The algorithm employs a characteristic
____________________________________
978-1-4799-8353-7 /15/$31.00 ©2015 IEEE