Research on Blind Source Separation Based on Natural Gradient Algorithm
SUN Shou-yu1 ,ZHENG Jun-li1,WU De-wei2 (1. Department of Electronics Engineering, Tsinghua University, Beijing 100084, China;2The Telecommunication Engineering Institute, Air Force Engineering University, Xi′an,Shaanxi 710077, China)
Blind source separation attempts to recover unknown independent sources from a given set of observed mixtures. The natural gradient algorithm is introduced in this paper, and it is a very important approach for blind source separation. We have examined the algorithm with communication signals and natural speech signals by different activation functions respectively. Simulation results demonstrate the algorithm can effectively separate the two kinds of random mixing signals.