基于混沌神经网络的二维局部均值估计自适应滤波

Two-dimension Local-mean Estimator Adaptive Filter Based on Chaotic Neural Network

为了提高自适应滤波的精度和收敛速度,提出了一种基于混沌神经网络的二维均值估计(LME)自适应滤波算法,在传统的二维LME自适应滤波方案中引入了混沌神经网络控制机制,用混沌神经网络自适应滤波器代替LME中的LMS自适应滤波算法,应用混沌神经网络估计局部期望输出进行滤波。仿真结果表明,该局部均值估计滤波器当输入信号为均值不为0且变化较大时,输出信号仍能较好地实现对输入信号的跟踪,获得了原始信号的主要特性,从均方误差曲面来看,算法具有较快的收敛速度和较高的滤波精度。

In order to enhance precision and convergence speed of adaptive filer,a two-dimension Local-mean Estimatior（LME） adaptive filter based on Chaotic Neural Network（CNN） was proposed in which CNN adaptive filter took the place of LMS algorithm in traditional LME adaptive filter to estimate local expected output. Simulation results indicate the ouput signal of the LME adaptive filter based on CNN can track input signal well and acquire the main character of orginal signal even input signal has non-zero mean and change greatly, on the other hand, the proposed algorithm has fast convergence speed and high filter precision.