机载雷达杂波抑制的预条件多级维纳滤波器算法

Preconditioned Multistage Wiener Filter Algorithm for Clutter Mitigation of Airborne Radar

在机载雷达抑制地杂波的应用中,降阶空时自适应处理是目前研究的重点,而在各种降阶算法中以Goldstein和Reed等人提出的多级维纳滤波器(MWF)算法最具代表性。本文首先将MWF算法由广义旁瓣对消结构推广到直接处理结构,在此基础之上引入预条件处理以改善杂波和噪声协方差矩阵的条件数,提出了预条件多级维纳滤波器(PMWF)算法,并给出了算法的预条件共轭梯度法实现。利用杂波和噪声协方差矩阵的分块Toeplitz结构,可使PMWF算法的计算量降低为O(NKlog_2NK)。仿真实验表明,PMWF算法与经典的MWF算法相比,当处理阶数相同时,其信杂噪比改善有明显的提高,且以主杂波附近的改善尤为显著。

Space-time adaptive processing is an effective technology for clutter mitigation of airborne radar, and the rank-reduced STAP methods are the emphasis of current researches. In these rank-reduced STAP methods, the multistage Wiener filter which was proposed by Goldstein and Reed has been shown to provide superior performance. In this paper,we first extend MWF from the generalized sidelobe canceller structure to the direct-form process structure. Then the preconditional process is utilized to improve the condition number of clutter covariance matrix. The preconditioned MWF method is proposed ,which can be realized by the form of preconditioned conjugate gradient method. The computation of PMWF method is only O（ NKlog2NK） owing to the block Toeplitz structure of clutter covarianee matrix. Simulations show that the SINR improvement of PMWF method is much higher than that of the MWF method. And the improvement is especially remarkable in the mainlobe region.