基于时间可预测性的差分搜索盲信号分离算法

Blind signal separation algorithm based on temporal predictability and differential search algorithm

针对基于仿生智能优化的盲信号分离算法计算量偏大的问题,提出了一种新的基于差分搜索的盲信号分离算法。采用信号在时间上的可预测性度量作为目标函数,使用差分搜索算法对目标函数进行优化求解。利用去相关消源方法从混合信号中去除每次分离出的源信号成分,通过逐次分离最终实现对所有源信号的成功恢复。仿真实验表明,所提算法可以有效实现对混合信号的盲分离。与其他算法相比,该算法在保证了更高分离精度的同时,具有更低的运算量。

A novel blind signal separation algorithm based on differential search was proposed for solving the high cal- culated amount problem in blind signal separation algorithm based on bio-inspired optimization. The temporal predict- ability of signal was used as the objective function and the differential search algorithm was used for solving it. The source signal component separated was wiped off using deflation method and all the source signals could be recovered successfully by repeating the separation process. Simulation results show that the algorithm can achieve blind separation from mixed signals efficiently with very high separation precision and very low computing time.