基于迭代算法的mimo雷达多目标参数估计方法研究.pdf

南京航空航天大学 硕士学位论文 基于迭代算法的MIMO雷达多目标参数估计方法研究 姓名：马晓颖 申请学位级别：硕士 专业：测试计量技术及仪器 指导教师：刘文波 2011-03 南京航空航天大学硕士学位论文 i 摘 要 多输入多输出(Multiple input Multiple output， MIMO)雷达是近年来提出的一种新体制雷达， 利用波形分集技术，可获得多于雷达真实阵元的虚拟阵元，提高了雷达的波达方向估计和目标 分辨性能本文主要研究 MIMO 雷达多目标参数估计问题，完成的工作如下 本文提出一种基于峰值临域搜索的 MVDR 谱 MIMO 雷达参数估计算法，通过搜索频谱峰 值临域内包含所有目标方位角的点组，有效地克服了传统方法普遍存在的谱线不匹配现象，缓 解谱线幅值衰减提高了目标参数估计的精度 本文提出一种基于期望最大化（EM）的 MIMO 雷达多目标参数估计算法该算法在每次 迭代过程中首先计算扩展后的接收信号矩阵期望值，随后利用极大似然方法估计每个目标的参 数实验表明，该算法的估计精度和角度分辨率有明显的提高 随后，将 EM 算法推广至匹配滤波后的双基 MIMO 雷达信号模型，并结合信息论准则，提 出同时估计目标个数及参数的方法。

该算法首先设定目标个数，利用 EM 方法估计目标参数， 随后根据特定的信息论准则设定模型阶数惩罚函数来判定设定的目标个数是否为真实目标个 数实验表明，与利用旋转不变技术估计信号参数（ESPRIT）方法相比，该算法提高了目标估 值精度，当信噪比大于- 2dB 时，可准确的估计目标个数 最后提出一种基于迭代思想的双基 MIMO 双边 APES 参数估计方法将传统的单边约束 APES 方法推广到匹配滤波后的信号模型，结合迭代思想提出了一种双边约束方法，依次估计 波达方位角和离波方位角理论分析及仿真实验验证了本文算法有效的降低了运算复杂度和样 本需求数 关键词：MIMO 雷达，参数估计，峰值临域搜索，期望最大化算法，信息论准则，双边 APES 算法 基于迭代思想的 MIMO 雷达多目标参数估计方法研究 ii ABSTRACT Multiple Input Multiple Output (MIMO) radar is a new radar system that has been proposed recently. By applications of spatial diversity, MIMO radar can obtain more virtual elements and improve the estimation performance of the location of the targets compared to the traditional phased array radar. This dissertation is addressed on the targets parameters estimation in the MIMO radar system, the main work of this dissertation can be summarized as the following. A novel MIMO radar parameters estimation algorithm based on peak frequencies local searching of MVDR spectrum is proposed. Comparing with the tradition algorithms, it can effectively overcome the spectrum mismatch and amplitude attenuation by searching the adjacent points of peak value, which include all the targets’ azimuth angle. Simulation results show efficiency of the new algorithm. A novel parameters estimation algorithm based on expectation maximization (EM) method for the multiple targets in MIMO radar system is proposed. In each iteration, the proposed algorithm gets the expectations of the received data matrix and then estimates the parameters of every target by the maximum likelihood (ML) method. Experimental results illustrate that the proposed algorithm can efficiently improve the estimation accuracy and the angular resolution . For bistatic MIMO radar, a joint estimation for the parameter and number of targets is proposed by improved the EM method, combined with the information- theoretic criteria, the target parameter is estimated by the EM method and then the targets number is judged by the penalty function derived from the selected information- theoretic criteria. Compared with the ESPRIT algorithm, the proposed algorithm can get the accurate number of targets when the SNR is larger than - 2dB and achieves a visible performance improvement of the estimation accuracy. A novel bi- boundary APES method is proposed for the bistatic MIMO radar system. The proposed algorithm adjust the traditional APES method to the targets return models after matching filter of bistatic MIMO radar and estimate DOA and DOD sequentially, combined with iterative method, the bi- boundary limitation method is derived. Theory analysis and e xperimental results illustrate that the proposed algorithm can significantly reduce the e computational load and training requirement. Key Words ： MIMO Radar, Parameter Estimation, peak frequencies local searching, EM, information- theoretic criteria, bi- boundary APES 南京航空航天大学硕士学位论文 v 图 清 单 图 2.1 单基地 MIMO 雷达示意图............................................................................................7 图 2.2 空间谱曲线比较图......................................................................................................10 图 2.3 方位角随步长变化图..................................................................................................11 图 2.4 计算量和估计误差随步长变化趋势图.........................................................................11 图 3.1 收、发共位的均匀线阵合成的等效一维虚拟线阵.......................................................14 图 3.2 收、发不共位的两平行均匀线阵合成的等效一维虚拟线阵.........................................15 图 3.3 收、发不共位的两垂直均匀线阵合成的等效二维虚拟面阵.........................................15 图 3.4 线阵和面阵合成等效虚拟立体阵................................................................................16 图 3.5 收、发共位不等距线阵合成的等效一维虚拟线阵.......................................................16 图 3.6 空间谱曲线比较图.....................................................................................................20 图 3.7 目标数未知时的空间谱曲线.......................................................................................21 图 3.8 波达方向估计 RMSE 曲线性能对比图.........................................................................22 图 3.9 算法收敛性能结果图.................................................................................................22 图 4.1 双基 MIMO 雷达示意图.............................................................................................23 图 4.2 MIMO 雷达接收端处理流程示意图.............................................................。