非均匀环境下基于知识辅助的扩展目标Wald检测器

摘要

This paper deals with the problem of detecting the moving range-extended target in the distributed MIMO radar.Firstly,the interference covariance matrices corresponding to different transmit-receive (Tx-Rx) antennas are modeled as random matrices which express nonhomogeneous environments.Then a knowledge-aided model which makes these random matrices share a prior covariance matrix structure is built to simulate the characteristics of clutter and noise in nonhomogeneous environments.On this basis,we design a new knowledge-aided Wald (KA-Wald) detector.Simulation results show that the proposed detector possesses a better detection performance compared with the traditional Wald detector.And relative to the knowledge-aided generalized likelihood ratio test (KA-GLRT) detector,the proposed KA-Wald detector has a similar detection performance but a higher efficiency.%针对分布式多输人多输出(Multiple Input Multiple Output,MIMO)雷达中运动扩展目标的检测问题,本文首先假设每个发射-接收天线组的干扰信号协方差矩阵为互不相同的随机矩阵,以模拟实际的非均匀工作环境.然后引入知识辅助模型,建立先验信息矩阵,描述非均匀环境下的干扰信号特性,其中所有发射-接收天线组的干扰协方差矩阵服从以先验信息矩阵为基础的逆Wishart分布.在此基础上,设计了一种基于知识辅助的Wald(KA-Wald)检测器.仿真实验表明,在小样本的情况下,本文设计的KA-Wald检测器在检测性能上优于传统Wald检测器.而与已有的基于知识辅助的广义似然比检验(KA-GLRT)检测器相比,检测性能相近,但是计算效率更高.