TDCS压缩感知观测矩阵构造方法  

Constructive method of CS measurement matrix in TDCS

作　　者：李楠[1] 任清华[1] 苏玉泽 

机构地区：[1]空军工程大学信息与导航学院

出　　处：《计算机工程与设计》2017年第1期7-11,共5页Computer Engineering and Design

基　　金：国家自然科学基金项目(61401499)

摘　　要：针对TDCS宽带压缩频谱感知使用随机性压缩感知观测矩阵结构复杂且不确定的问题,提出一种确定性嵌入式混沌序列-循环Toeplitz结构压缩感知观测矩阵构造方法。选用初值为0.2的四阶Chebyshev型嵌入Logistic型,形成嵌入式混沌序列,产生相关性渐进最优的确定序列,进行循环Toeplitz结构形成观测矩阵。仿真结果表明,嵌入式混沌序列-循环Toeplitz结构观测矩阵较随机高斯、随机贝努利观测矩阵有更好的稳定性和RIP特性,在相同稀疏度下具有更小的重构误差,更易精确重构。Random compressed sensing （CS） measurement matrix structure used in TDCS wideband compressed spectrum sensing is complex and uncertain, to solve the problem, a deterministic CS measurement matrix construction method was put forward, namely embedded chaos sequence-cycle Toeplitz structure. The fourth-order Chebyshev type whose initial value was 0.2 was em- bedded into Logistic type to form embedded chaotic sequence. The determination sequence with optimal correlation was produced. Deterministic CS measurement matrix was formed by cycle-Toeplitz structure. Simulation results show that embedded chaos sequence-cycle Toeplitz structure measurement matrix has better RIP speciality and stability compared to random Gaussian and random Bernoulli. In the same sparse degree, it has smaller reconstruction error and can be reconstructed more accurately.

关 键 词：TDCS频谱感知 压缩感知观测矩阵 嵌入式混沌序列 循环Toeplitz结构 

分 类 号：TN92[电子电信—通信与信息系统]