基于信息论的MIMO压缩感知雷达测量矩阵优化

发布时间：2018-05-09 05:08

本文选题：压缩感知雷达 + 测量矩阵优化　；参考：《西安电子科技大学》2014年硕士论文

【摘要】：超宽带、高分辨、多通道及多目标已成为目前雷达处理技术的主要发展趋势,由此带来了雷达信号处理过程中运算量巨大,数据存储量、传输量急剧上升,实时处理困难等问题,受传统奈奎斯特(Nyquist)采样定理的限制,大数据处理成为传统雷达的一个难以突破的瓶颈。压缩感知雷达由于实现了对雷达数据的直接采样,降低了系统的采样速率和系统复杂度,从而提高系统的分辨率,其在雷达领域有着广阔的应用前景。测量矩阵与稀疏变换基的随机不相关特性是压缩感知雷达正确恢复信号的基础,测量矩阵的好坏既直接关系到信息采集和信息传递的实现,又关系到能否从低维投影中正确恢复原始信号。传统的随机测量矩阵的矩阵存储元素容量巨大,计算复杂度高,并且不是最优的测量矩阵。针对这种情况,为了改善压缩感知雷达目标参数检测和成像性能,本论文研究了MIMO压缩感知雷达测量矩阵优化问题,仿真分析验证了已有的基于最大信噪比准则的测量矩阵优化方法,并提出了一种基于互信息准则的测量矩阵优化方法。文中首先建立了高斯伯努利随机信号模型下MIMO压缩感知雷达成像场景,然后根据信息论和凸优化的理论知识推导了最大互信息与最小互相关系数测量矩阵优化的目标函数和约束条件,并运用优化工具包优化求得测量矩阵,最后用优化所得的测量矩阵观测数据,以SLIM算法重构目标信号实现目标检测和成像。计算机仿真实验验证了基于互信息准则的测量矩阵优化方法能够使得感知矩阵列与列之间的互相关系数更小,从而得到更好的测量矩阵。该方法下的测量矩阵克服了现有的随机性测量矩阵实现困难和确定性测量矩阵恢复精度低的问题,能够有效提高目标重构性能,降低目标重构误差,提高可检测目标性能和成像性能。
[Abstract]:Ultra-wideband, high-resolution, multi-channel and multi-target have become the main development trend of radar processing technology. Due to the limitation of Nyquist sampling theorem, big data processing has become a bottleneck of traditional radar. Compressed sensing radar has a wide application prospect in radar field because it realizes direct sampling of radar data, reduces the sampling rate and complexity of the system, and improves the resolution of the system. The random discorrelation between the measurement matrix and the sparse transform basis is the basis for the correct recovery of the signal in the compressed sensing radar. The quality of the measurement matrix is directly related to the realization of information collection and information transfer. It is also related to whether the original signal can be correctly recovered from the low dimensional projection. The matrix of traditional random measurement matrix has large storage capacity, high computational complexity, and is not the optimal measurement matrix. In order to improve the target parameter detection and imaging performance of compressed sensing radar, this paper studies the optimization of MIMO compressed sensing radar measurement matrix. Simulation results verify the existing measurement matrix optimization methods based on the maximum signal-to-noise ratio (SNR) criterion, and propose a measurement matrix optimization method based on mutual information criterion. In this paper, firstly, the MIMO compression sensing radar imaging scene based on Gao Si Bernoulli random signal model is established. Then, according to the information theory and the theory knowledge of convex optimization, the objective function and constraint conditions of the measurement matrix optimization of the maximum mutual information and the minimum correlation number are deduced, and the measurement matrix is obtained by optimizing the optimization tool kit. Finally, the target detection and imaging are realized by using SLIM algorithm to reconstruct the target signal by using the optimized measurement matrix observation data. The computer simulation results show that the method based on mutual information criterion can make the number of correlation between the columns of perceptual matrix and the columns smaller and obtain a better measurement matrix. The measurement matrix based on this method overcomes the difficulties in the implementation of the existing random measurement matrix and the low recovery accuracy of the deterministic measurement matrix. It can effectively improve the performance of target reconstruction and reduce the error of target reconstruction. Improve detectable target performance and imaging performance.
【学位授予单位】：西安电子科技大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN957.51

【相似文献】