摘要
提出一种对FIR滤波器和基于反馈RBF网络的非线性滤波器都适用的改进型NLMS(VS MNLMS)算法,并将其应用于线性和非线性自适应逆控制(AIC)系统的逆建模.该算法计算简单,容易实现,具有全局收敛性.数字仿真结果表明VS MNLMS能获得比其它四种变步长LMS算法更快的收敛速度、更小的稳态MSE,更好的鲁棒性,并使AIC系统具有良好的动静态性能,从而验证了本文提出的算法和非线性滤波器在AIC中的有效性.
An Improved NLMS algorithm （VS MNLMS） for both FIR filter and feedback RBF Network based nonlinear adaptive filter is presented in this paper, the proposed filter and algorithm are applied to the inverse model in both linear and nonlinear adaptive inverse control （AIC） system. This algorithm is simple, easy to implement and without local minimums. Practical simulation results show that this algorithm can achieve quicker speed of convergence, lower misadjustment error and better robustness than other four variable step-size LMS algorithms do, and make the AIC system to exhibit excellent dynamic and static performances. So it is verified that the proposed algorithm and RBFNN based nonlinear filter are efficient to adaptive inverse control.
出处
《小型微型计算机系统》
CSCD
北大核心
2008年第11期2171-2176,共6页
Journal of Chinese Computer Systems
基金
国家自然科学基金项目(60474041)资助
国家"八六三"项目(2004AA001032)资助
