摘要
本文提出一种新的自适应滤波器——Gamma 滤波器及其基于最小均方误差准则的训练方法。首先建立了最小均方误差准则下 Gamma 滤波器的 Wiener-Hopf 方程,然后分别提出3种训练算法:(1)高斯-牛顿算法;(2)确定性梯度算法;(3)LMS算法。最后给出Gamma 滤波器在随机信号预测中的应用和仿真结果。
A new adaptive filter--the Gamma filter and its minimum-mean-
suare-error criterion-based training methods are presented. We first derive the Wiener-Hopf equations for the Gamma filter using the minimum-mean-square-er-ror criterion, and then develop three methods for training the Gamma filter: (1) Gauss-Newton method; (2) deterministic gradient descent method; (3) the LMS method. Finally the application of the Gamma filter to random signal prediction and the corresponding simulation results are given.
出处
《测试技术学报》
1997年第4期1-6,共6页
Journal of Test and Measurement Technology
关键词
滤波器
最小均方误差
自适应滤波器
the Gamma filter
minimum-mean-square-error
the wiener-hopf e-quations
