【摘要】 针对分段正交匹配追踪(St OMP)算法需要信号的稀疏度作为先验信息且重构精度较低的特点,提出一种稀疏度自适应分段正交匹配追踪算法。首先,通过对观测矩阵与初始残差相乘所得的残余相关性向量进行离散余弦变换,估算出支撑集所要扩充的最大原子数;其次,采用与抽样率成正相关的因子对较大的阈值参数进行适当修正,并对通过设定阈值所选取的原子进行优化处理;最后在St OMP算法的框架下采用变步长的方法实现稀疏度的逼近和信号的精确重构。仿真结果表明:本文所提出的算法对信号的稀疏度具有很好的自适应特性,并且在保持了较低重构复杂度的同时具有更稳定的重构质量。更多还原

【Abstract】 An improved stagewise orthogonal matching pursuit(St OMP) algorithm was proposed considering that the algorithm needs signal sparsity as the prior knowledge and has a relatively poorer reconstruction performance. Firstly, discreet cosine transform was applied to the vector of residual correlations to estimate the maximum number of atom needed by the support set. Then the large threshold parameter was adjusted by a factor which is positively correlated with the sampling rate and the atoms chosen by setting a threshold value was optimized. Finally, the close approach of signal sparsity and precise reconstruction of the signal were realized with variable step size within the frame of St OMP. The results show that the proposed algorithm has good adaptability without prior information of the sparsity and this algorithm not only keeps the low reconstruction complexity but also shows better and more stable reconstruction quality than the original St OMP algorithm.更多还原