Research On Reduced-Dimension Algorithms In Array Paremeter Estimation

Direction of arrival(DOA) estimation is the key point of array signal processing research, and multiple parameter estimation has become the study hotspot with the development of the theory of array signal processing. The multiple parameter estimation including two dimensional DOA(2D-DOA) estimation, joint estimation of angle and frequency, joint estimation of angle and polarization estimation, joint estimation of direction of departure(DOD) and DOA in multiple-input multiple-output(MIMO) radar. To expand common spectral peak searching algorithm to multiple parameter estimation, which can realize multiple parameter estimation based on multi-dimensional spectrum peak search algorithms. However, the computation complexities of multi-dimensional spectrum peak search algorithms are very high, which limits the application of this kind of algorithms. So, the research on reduced-dimension algorithms in array paremeters estimation has great significance in theory and application. The main work of this paper is summarized as follows:(1) A 2D-DOA estimation for uniform rectangular array using reduced-dimension multiple signal classification(RD-MUSIC) algorithm is investigated. Compared to two dimensional MUSIC(2D-MUSIC), the RD-MUSIC algorithm only requires one dimensional local searching by dimension reduction transformation, which avoids two dimensional global searching and has lower conputation complexity. The simlution results show that the angle estimation of RD-MUSIC algorithm is very colse to 2D-MUSIC algorithm.(2) A 2D-DOA estimation for uniform rectangular array using reduced-dimension propagator method(RD-PM) algorithm is propoesd. The proposed algorithm gets initial estimation with rotation invariance of propagator matrix, then obatins more accurate estimation through one dimensional local searching. The proposed algorithm, which requries no eigen value decomposition of the cross correlation matrix of the receive data and avoid two dimensional global peak searching, has low complexity. Also, the parameters are automatically paired.(3) We consider the DOD and DOA joint estimation problem for non-uniform arrays MIMO radar, and propose a blind DOD and DOA estimation algorithm using RD-PM. The proposed algorithm simplifies the two-dimensional global searching in two-dimensional propagator method(2D-PM) to one-dimensional local searching, and achieves automatically paired two-dimensional angle estimation. Meanwhile, the proposed algorithm requires no eigen-value decomposition of the covariance matrix of the receive data and has a low complexity than the generalized ESPRIT algorithm. The angle estimation performance is better than that of the ESPRIT-like algorithm, the PM-like algorithm and the generalized ESPRIT algorithm, also is almost the same to 2D-PM algorithm.(4) A joint estimation of DOD, DOA and polarization information for bistatic polarimetric MIMO radar using RD-MUSIC is proposed. The proposed algorithm obatins initial estimation through estimation of signal parameters via rotational invariance techniques(ESPRIT) algorithm, then uses dimension reduction transformation and gets more accuracy estimaiton. The proposed algorithm requires no four dimensional peak searching and has much lower computation complexity than four dimensional MUSIC(4D-MUSIC) algorithm. Also, the DOD, DOA and polarization parameters are automatically paired.