摘要
本文着重研究了自适应滤波器的重要实现形式——递推最小二乘算法(RLS)的原理,分析了RLS算法在应用中的优点及存在问题。为解决RLS算法收敛速度和稳态误差的矛盾及系统在趋于平稳时跟踪效果差的问题,本文从实现可变遗忘因子和增加自扰动项两个方面介绍了RLS算法的几种改进方法。并将它们应用于复杂电磁环境、强干扰背景下的信号分离中去。通过仿真实验,对RLS算法及其两种改进方法在信号分离中的效果进行了比较,得出可变遗忘因子RLS算法在收敛速度和分离信号的准确性上都具有较好的性能。
This paper focuses on an important form of adaptive filter——recursive least squares algorithm（RLS）,analysis of the merits and problems of the RLS algorithm in the application.In order to solve the contradiction between convergence speed and steady-state error and system to track results in stabilizing the problem of poor,in this paper,achieving a variable forgetting factor and increased disturbance from RLS algorithm presented of several improved methods in the two aspects.And apply them in signal separation with the context of strong interference and complex electromagnetic environment.Then by simulation,taking the RLS algorithm and two improved methods of signal separation results in a comparison,we can take the conclusion that variable forgetting factor RLS algorithm in convergence speed and accuracy of signal separation,both with good performance.
出处
《电子测试》
2012年第1期23-27,共5页
Electronic Test
关键词
递推最小二乘算法
信号分离
可变遗忘因子
误差分析
recursive least squares
signal separation
variable forgetting factor
error analysis
