基于信息论的盲源信号分离

Blind source separation based on information theory

从混合观测数据向量中恢复出不可直接观测的各个源信号是阵列处理和数据分析的典型问题.独立分量分析是解决这一类问题的新技术,而基于信息论方法的分离技术是独立分量算法中最常用的分离算法.基于信息论算法中主流的FastICA算法和自然梯度优化算法,使用几组不同的信号进行分离,从理论分析和仿真结果表明了FastICA算法的优越性.

Recovering the unobserved source signals from their mixtures is a typical problem in array processing and analysis. Independent component analysis （ICA） is a new method to solve this problem. The most common way in independent component analysis is the separation based on information theory. FastlCA algorithm and nature step algorithm are the main way in it. Some groups of signals were separated. The analysis and simulations suggest that the FastlCA algorithm is the best way.