摘要
针对传统压缩感知理论无法实现自适应压缩测量的不足,介绍了基于贝叶斯估计的压缩感知重构算法;建立了基于重构信号微分熵的自适应评价指标;构建了基于分块测量方式的自适应压缩测量算法。仿真分析了该算法用于随机阶跃信号的测量效果,实验结果表明:建立的自适应评价指标与被测信号重构误差具有相似的变化趋势,能有效反映随机测量进程,可以实现自适应压缩测量。正是在BCS理论的基础上,提出了一种根据以获取信息的不确定度,进而自适应实现压缩测量的算法,该算法运算简单,适用于实时信号的在线压缩测量。
To overcome the challenge of traditional Compressive Sensing without adaptive measurement ability, the theory of Bayesian Compressive Sensing is briefly introduced. An evaluation index based on differential entropy of estimated signal is devised and the adaptive compressive measurement procedure without any prior information about the measured signals is presented in block manner. Numerical simulations on random step signal verify that the adaptive algorithm has good performance. This algorithm offers great potential for adaptive compressive measuring.
出处
《计算机工程与应用》
CSCD
2013年第9期200-202,217,共4页
Computer Engineering and Applications
关键词
压缩感知
贝叶斯估计
自适应测量
微分熵
重构误差
Compressive Sensing（CS）
Bayesian estimation
adaptive measurement
differential entropy
reconstruction error
