New Algorithms for Blind Source Separation
A novel measure of signal temporal predictability is proposed, based on which, the new algorithm is more simplified and less time-consuming than the method based on ICA and high-order statistic, and it requires minimal assumptions regarding the probability density functions of the source signals. The new algorithms in this work are proposed in both batch mode and online mode. Simulations show that the new algorithm for batch mode has the best precision, and its convergence, stability and robustness are quite well for online mode.
Zuyuan Yang Anan Deng
School of Electrics & Information Engineering South China University of Technology Guangzhou, P.R.China 510641
国际会议
南宁
英文
2007-07-20（万方平台首次上网日期，不代表论文的发表时间）