基于极化分集技术与随机矩阵理论的MIMO雷达目标检测方法

发布时间：2018-05-08 09:07

本文选题：MIMO雷达 + 目标检测　；参考：《吉林大学》2014年硕士论文

【摘要】：多输入多输出（MIMO）雷达在目标检测、参数估计和目标识别等领域具有诸多优势。由于采用了分集的发射信号，MIMO雷达可显著提高目标的检测性能。本文首先针对基于极化分集技术的MIMO雷达目标检测问题展开了深入研究。传统MIMO雷达主要采用空间分集技术，而极化分集技术的利用使雷达在小目标和杂波背景下具有更强的检测能力。不同的极化方式会带来检测性能的较大差异，通过发射端极化波形的优化设计，可明显改善MIMO雷达的检测性能。进而，本文研究了基于随机矩阵理论（RMT）的MIMO雷达目标检测问题。目前MIMO雷达目标检测大多是在采样数远大于阵元数的假设前提下进行的，当样本数不充足时导致性能降低。随机矩阵理论为MIMO雷达信号处理提供了一个便利的工具，在噪声方差和目标散射矩阵未知的环境下，基于随机矩阵渐进谱理论（AST）的方法可实现双基地MIMO雷达目标的盲检测。本论文的研究工作得到国家自然基金项目“基于极化分集的MIMO雷达参数联合估计与目标定位”（项目编号：61071140）和“基于大维随机矩阵的MIMO雷达稳健目标检测与估计”（项目编号：61371158）的资助。本文主要研究工作如下：在传统分布式MIMO雷达模型的基础上，对杂波背景下基于极化分集的MIMO雷达目标检测问题进行了研究。建立了极化MIMO雷达目标检测的信号模型，提出了一种基于Jones矢量的极化MIMO雷达检测算法。该方法利用发射天线波形极化的多重搜索，实现了极化波形的优化。仿真结果验证了算法的有效性，与水平、垂直、正交极化方式相比，，该算法改善了目标检测的性能。为降低多重搜索的复杂度，本文进一步提出一种基于萤火虫群优化（GSO）的MIMO雷达目标检测算法。该方法以检测概率最大为目标函数进行发射极化波形的选择，利用GSO算法进行多维并行搜索，通过并行处理数据，同时优化多个极化参数，解决了难以实现的多重嵌套搜索问题。对算法的仿真结果表明，基于GSO的MIMO雷达目标检测算法改善了检测性能，减少了计算量。上述MIMO雷达检测方法尽管提高了目标检测性能，然而，当采样数不足或采样数与阵元数接近时其性能将会下降。鉴于此，本文从双基地MIMO雷达模型出发，在采样数与收发阵元数的乘积接近的情况下，提出一种基于随机矩阵理论的MIMO雷达目标检测算法。该方法在目标散射信息与定位信息及噪声方差未知的情况下，利用随机矩阵理论实现了目标的盲检测。该算法对先验要求大大放松，对噪声变化不敏感，实现了大阵列情况下MIMO雷达的稳健目标检测。
[Abstract]:Multi-input multiple-output MIMO-radar has many advantages in target detection, parameter estimation and target recognition. Because of the diversity of transmit signal MIMO radar can significantly improve the detection performance of the target. In this paper, the problem of MIMO radar target detection based on polarization diversity is studied. The traditional MIMO radar mainly uses space diversity technology, but the use of polarization diversity technology makes radar have stronger detection ability in the background of small target and clutter. Different polarization modes will lead to great differences in detection performance. The detection performance of MIMO radar can be improved obviously by optimizing the polarization waveform of the transmitter. Furthermore, the problem of MIMO radar target detection based on random matrix theory is studied in this paper. At present, MIMO radar target detection is mostly based on the assumption that the number of samples is much larger than the number of array elements, and the performance is degraded when the number of samples is not sufficient. Stochastic matrix theory provides a convenient tool for signal processing of MIMO radar. Under the condition of unknown noise variance and target scattering matrix, blind detection of bistatic MIMO radar targets can be realized based on stochastic matrix asymptotic spectrum theory. In this paper, the National Fund for Nature Project "Joint estimation and Target location of MIMO Radar parameters based on polarization Diversity" (Project No.: 61071140) and "robust Target Detection and estimation of MIMO Radar based on large Dimension Random Matrix" "(item No. 61371158). The main work of this paper is as follows: Based on the traditional distributed MIMO radar model, the problem of MIMO radar target detection based on polarization diversity in clutter background is studied. The signal model of polarimetric MIMO radar target detection is established, and a polarimetric MIMO radar detection algorithm based on Jones vector is proposed. In this method, the polarization waveform of transmitting antenna is optimized by multiple search. Simulation results verify the effectiveness of the algorithm. Compared with horizontal, vertical and orthogonal polarization, the algorithm improves the performance of target detection. In order to reduce the complexity of multiple search, a MIMO radar target detection algorithm based on firefly swarm optimization (GSO) is proposed in this paper. In this method, the maximum detection probability is taken as the objective function to select the transmitted polarization waveform, and the multi-dimensional parallel search is carried out by using GSO algorithm, and the data is processed in parallel, and several polarization parameters are optimized at the same time. It solves the problem of multi-nested search which is difficult to implement. The simulation results show that the MIMO radar target detection algorithm based on GSO improves the detection performance and reduces the computational complexity. Although the above MIMO radar detection method improves the performance of target detection, its performance will decline when the number of samples is insufficient or the number of samples is close to the number of array elements. In this paper, based on the bistatic MIMO radar model, a MIMO radar target detection algorithm based on random matrix theory is proposed under the condition that the product of sampling number and transceiver element number is close to each other. Under the condition that the scattering information and location information and the variance of noise are unknown, the blind detection of target is realized by using the stochastic matrix theory. The algorithm is not sensitive to noise changes and greatly relaxes the priori requirements, and realizes robust target detection in large array MIMO radar.
【学位授予单位】：吉林大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN957.51

【参考文献】