摘要
该文利用复数稀疏信号的时域相互关系提出一种新的稀疏贝叶斯算法(CTSBL)。该算法利用复数信号的实部与虚部分量具有相同的稀疏结构的特点,提升估计信号的稀疏程度。同时将多个测量信号间的内部结构信息引入到了信号恢复中,使原始的多测量稀疏信号恢复问题转变为单测量块稀疏信号恢复问题,使恢复性能得到了提升。理论分析和仿真结果证明,提出的CTSBL算法相较于目前的针对复数信号的多测量矢量贝叶斯压缩感知(CMTBCS)算法和块正交匹配追踪算法(BOMP)在估计精度上具有更好的性能。
An effective Sparse Bayesian Learning algorithm exploiting Complex sparse Temporal correlation（CTSBL） is proposed in this paper, which is used to recover sparse complex signal. By exploiting the fact that the real and imaginary components of a complex value share the same sparsity pattern, it can improve the sparsity of the estimated signal. A multitask sparse signal recovery issue is transformed to a block sparse signal recovery issue of a single measurement by taking full advantage of the internal structure information among the multiple measurement vector signals. The experiments show that the proposed algorithm CTSBL achieves better recovery performance compared with the existing Complex Multi Task Bayesian Compressive Sensing（CMTBCS） algorithm and BOMP algorithm.
出处
《电子与信息学报》
EI
CSCD
北大核心
2016年第6期1419-1423,共5页
Journal of Electronics & Information Technology
基金
国家自然科学基金(61571148)
中国博士后科学基金(2014M550182)
黑龙江省博士后特别资助(LBH-TZ0410)
哈尔滨市科技创新人才资助课题(2013RFXXJ016)
中国博士后特别资助(2015T80328)~~
关键词
压缩感知
稀疏信号恢复
多矢量测量模型
块稀疏贝叶斯
Compressive Sensing（CS）
Sparse signal recovery
Multitask measurement vector model
Block sparse Bayes
