基于块稀疏矩阵恢复的MIMO雷达扩展目标高分辨成像算法  

Extended target high resolution imaging algorithm for MIMO radar based on block sparse matrix recovery

作　　者：蒲涛 童宁宁[1] 冯为可 房亮 高晓阳 PU Tao;TONG Ningning;FENG Weike;FANG Liang;GAO Xiaoyang(Air and Missile Defense College, Airforce Engineering University, Xi’an 710051, China;Research Institute for Special Structures of Aeronautical Composites, Aivation Industry Corporation of China, Jinan 250031, China;Unit 93046 of the PLA, Qingdao 266111, China)

机构地区：[1]空军工程大学防空反导学院,陕西西安710051 [2]中国航空工业集团有限公司济南特种结构研究所,山东济南250031 [3]中国人民解放军93046部队,山东青岛266111

出　　处：《系统工程与电子技术》2021年第3期647-655,共9页Systems Engineering and Electronics

基　　金：国家自然科学基金(61901511,61701526);陕西省自然科学基金(2019JM-322);航空科学基金(ASFC-201918096001)资助课题。

摘　　要：针对现有多输入多输出(multiple input multiple output,MIMO)雷达稀疏恢复成像算法中存在的运算量大、对扩展目标成像质量低的问题,提出一种基于块稀疏矩阵恢复的MIMO雷达扩展目标高分辨成像算法,通过引入目标块稀疏特征,提高对空间扩展目标的成像质量。首先,通过构造距离向和方位向感知矩阵,建立目标散射系数估计的块稀疏矩阵恢复模型。然后,采用分块二维序列一阶负指数(sequential order one negative exponential,SOONE)函数对目标块稀疏特征进行提取。最后,利用梯度投影算法对块稀疏矩阵范数优化问题进行求解,在欠采样条件下得到目标高质量图像。相比于传统成像算法,所提算法可以在实现对扩展目标高分辨成像的同时,降低数据采样量,且具有较高的准确性、鲁棒性和较低的运算量。仿真实验验证了所提成像算法的有效性。In order to solve the problems of high computational complexity and low imaging quality for extended targets in existing multiple input multiple output(MIMO)radar sparse recovery imaging algorithms,a high resolution imaging algorithm for MIMO radar extended targets based on block sparse matrix recovery is proposed.By introducing the block sparse feature of target,the imaging quality of spatial extended target is improved.Firstly,by constructing the range and azimuth sensing matrices,the block sparse matrix restoration model of target scattering coefficient estimation is established.Then,the block sparse features of the target are extracted by sequential order one negative exponential(SOONE)function.Finally,the gradient projection algorithm is used to solve the norm optimization problem of block sparse matrix,and the high-quality image of the target is obtained under the condition of under sampling.Compared with the traditional imaging algorithms,the proposed algorithm can achieve high-resolution imaging of extended targets while reducing the amount of data sampling,and has higher accuracy,robustness and lower computational complexity.Simulation experiments verify the effectiveness of the proposed imaging algorithm.

关 键 词：多输入多输出雷达 高分辨成像 稀疏矩阵恢复 块稀疏 

分 类 号：TN957.51[电子电信—信号与信息处理]