摘要
变步长自适应数字滤波器的权值随信号和噪声的变化自动调整时可能出现不稳定。为解决这一问题,本文提出变步长自适应数字滤波器的权值收敛条件,该条件具有较少保守性,对输入信号的统计特性没有严格限制。本文证明自适应数字滤波器的稳定性可由其权值收敛条件确定。利用该权值收敛条件,可设计确保变步长自适应数字滤波器的稳定的变收敛因子。
Adaptive digital filters with variable step-size may become unstable when their weights are automati- cally adjusted with input signal and noise. To solve the problem, we establish and prove the convergence condi- tion of the weights of the adaptive digital filters. The condition is less conservative, and of no rigorous limitation to the characteristics of the input signal. We show that the stability of variable step-size adaptive digital filters can be determined by the convergence of the weights of the filters. The convergence condition can be used to design the time-varying, convergence factor, and ensure variable step-size adaptive digital filters to be stable.
出处
《通信学报》
EI
CSCD
北大核心
2001年第10期88-92,共5页
Journal on Communications
基金
国家自然科学基金资助项目(69971002)
关键词
自适应数字滤波器
收敛条件
稳定性
adaptive digital filters
convergence conditions
stability