基于近似消息传递算法的压缩感知雷达成像方法  

Compressed sensing radar imaging using approximate message passing

作　　者：唐琳[1] 焦淑红[1] 

机构地区：[1]哈尔滨工程大学信息与通信工程学院,黑龙江哈尔滨150001

出　　处：《智能系统学报》2015年第4期592-598,共7页CAAI Transactions on Intelligent Systems

基　　金：国家自然科学基金资助项目(61201410)

摘　　要：针对大观测矩阵引起的压缩感知雷达成像高计算量的问题,提出了一种基于近似观测模型的近似消息传递算法,通过引入逆线调频变标算子,压缩感知雷达成像中的大观测矩阵能够得到很好的近似,从而能够有效地运用现有的成熟的去耦合措施来降低压缩感知雷达成像的计算量;同时,通过引入近似的消息传递,新算法获得了不错的收敛速度。理论分析和仿真实验表明,新算法能够实现非完全观测数据的压缩感知雷达成像,与现有的同类压缩感知雷达成像方法相比,其在保证不增加单步迭代计算量的同时,具有更高的收敛速度。In order to solve the great computational complexity in compressed sensing radar imaging due to a large observation matrix,an approximate observation model based approximate message passing algorithm is proposed in this paper. It uses inverse chirp scaling operator to approximate the large observation matrix in compressed sensing radar imaging,thereby effectively reducing the computational complexity by the current mature decoupling technology,and in the meantime,the approximate message passing is used to improve convergence rate. The theoretical analysis and simulation show that compared to the currently used compressed sensing radar imaging methods,the proposed method exhibits higher convergence rate while suffers same computational complexity in each iteration. It realizes compressed sensing radar imaging of non-complete observation data.

关 键 词：雷达成像 压缩感知 近似观测模型:消息传递 置信传播 

分 类 号：TN957.52[电子电信—信号与信息处理]