基于粒子群算法的自适应滤波器研究及应用

Researching and Application of Adaptive Filter Based on PSO Algorithm

针对神经网络自适应滤波器易于陷入局部极小的缺陷,采用抑制局部最优的粒子群算法优化神经网络的权系数,设计了基于改进粒子群算法训练的三层神经网络的自适应滤波器,并将其应用于自适应噪声抵消器.仿真表明,该系统与传统自适应噪声抵消系统相比具有很好的噪声抵消能力,信噪比大大提高.

In order to restrain the traditional adaptive filter based on neural network from trapping in local optimum, and an adaptive filter based on three - tier neural network was designed, hnpmving PSO algorithm was used to optimize the mutation operator according to standard derivation of swarm fit value to inhibit local optimum. An adaptive noise canceller was designed with an adaptive filter. The simulation shows that this system has better noise cancellation capability compared to the traditional adaptive noise canceller, and increases SNR. greatly.