基于OS-SASP算法的欠定盲源分离

针对基于稀疏分量分析的欠定盲源分离问题，提出一种基于优化支撑的稀疏度自适应子空间追踪(OS-SASP)算法.通过引入自适应思想，克服传统子空间追踪(SP)算法对稀疏度的依赖;同时在迭代开始之前通过离散余弦变换的能量集中特性确定最小支撑集的大小，对最小支撑集求并集获得优化支撑集，优化支撑集联合迭代过程中的候选集来定位最佳原子，提高源信号的恢复精度.仿真结果表明，OS-SASP算法在一维稀疏信号与语音信号的欠定盲源恢复过程中表现出良好的性能.

Underdetermined Blind Source Separation Based on OS-SASP Algorithm

A sparse adaptive subspace pursuit based on the optimal support(OS-SASP)algorithm was proposed to deal with the problem of underdetermined blind source separation based on sparse component analysis. By introducing the idea of self-adaptation， the dependence of the traditional subspace pursuit(SP)algorithm on sparsity was overcome. At the same time， the size of the minimum support set was determined by the energy concentration characteristic of discrete cosine transform before the start of iteration. Further， the optimal support set was obtained by calculating the union of the minimum support sets. And the combination of the optimal support set and the candidate set in the joint iteration was used to locate the best atom， so as to improve the source signal′s restore accuracy. The simulation results showed that the OS-SASP algorithm can achieve promising performance in the underdetermined blind source recovery of the one-dimensional sparse signals and speech signals.