基于卷积盲分离神经网络的过程信号去噪算法

Noise-cancellation algorithm of process signal based on blind signal sepration of convolutive mixtures neural network

研究了一种基于卷积混叠盲信号分离的简约神经网络算法,简称RCMNN。可使线性静态或动态传输通道中混叠的多源信号实现分离,从而可有效应用于传输通道中过程信号上的去噪,是控制工程中去噪的新方法。对多类不同信号组合仿真表明,该算法是有效的,网络性能是稳定的。

The model and algovithm ofconvolutive mixture for process signal, based on the blind separation of RCMNN neural network, are presented. The noise mixed on process signal by a linear and convolutive fashion in a stationary or nonstationary propagation passage can be canceled with this approach. This is a new noise-cancellation method in process control engineering.