盲信号分离的自适应算法研究

The Research on the Adaptive Algorithms of Blind Signal Separation

盲信号分离在信号处理界一直有着重要的地位,其目的就是要根据观测到的混合数据向量恢复出混合前的原始信号。论述LMS和RLS两种进行盲信号分离的自适应算法,并使用它们对合成数据进行实验,考察算法的特性和效果,并进行比较分析。结果表明:LMS算法与RLS算法相比,RLS算法的收敛性能更好一些,而RLS的稳定性则存在振荡问题,在实际应用中需要选择合适的方法和参数。

The separation of the blind signal is very important in the field of signal processing. The aim of BSS（Blind Signal Separation） is to recover the originalindependent source signals from their mixed signal vectors, by constructing a certain transform. The two adaptive algorithms of blind signal separation-LMS and RLS is introduced and make experiments using synthetic data and real data, then observe the characteristics and effects of the two algorithms , making analysis and comparison. The resalt shows that RLS algorithm has better convergence than LMS alyorithm and theres oscillation in RLS algon＇thm＇s stability. Appropriate parameters and methocls should be choosed in its practtcal application.