动态神经网络的超定盲信号分离

Dynamical Neural Network of Over-Determined BBS

在观察信号数目不小于源信号数目的条件下,提出了一种信号源数目未知或者信号源动态变化的盲信号分离算法。利用动态神经网络来控制输出信号的个数。经试验仿真,得到该算法分离的有效性。

Under this condition that here are more misxtures than sources,a dynamical neural network for BBS with an unkonw or dynamical changing source number is introduced.Using dynamical nural network to get the number of output source. And simulation results demonstrate the effectiveness of the proposed algorithm.