摘要
通过对传统盲源分离批处理EASI算法的分析,针对时变信道中通信信号的复数形式,以平滑窗的形式实现了批处理算法在时变混合模型下的应用。通过主分量分析的原理,进行噪声方差估计,改进传统独立分量分析算法消除噪声影响。通过仿真证明,在噪声较大的情况下性能优于传统算法。
Based on an analysis of EASI batch process algorithms for traditional blind source separation, a sliding window ICA algorithm is studied to deal with complex signals in the time variant mixing model. The characteristics of noise are estimated using Primary Component Analysis. The traditional Independent Component Analysis algorithm is improved to eliminate the effect of noise. The experiment results show that this algorithm is better than traditional ones with more noise present.
出处
《电子科技》
2007年第11期8-10,14,共4页
Electronic Science and Technology
关键词
盲源分离
平滑窗
独立分量分析
主分量分析
blind source separation
sliding window
independent component analysis
primary component analysis
