摘要
本文利用循环统计量讨论几乎周期时变滑动平均（APTV-MA）系统的辨识问题提出了基于奇数阶时变累量的参数估计的法方程方法，并导出了基于时变累量和循环累量的系统模型定阶方法．仿真实验说明了本文所提方法的性能．
This paper addresses the problem of identification of almost periodic time-varying moving average (APTVMA) systems using cyclic statistics. The nomal equation approches based on odd order time-varying cumulants are presentedfor parameters estimation. Model order detemination procedures based on time-varying and cyclic cumulants are also derived.Simulations are performed to illustrate performance of the presented methods.
出处
《电子学报》
EI
CAS
CSCD
北大核心
1999年第12期52-55,42,共5页
Acta Electronica Sinica
基金
国家自然科学基金
关键词
循环统计量
MA系统
系统辨识
cyclic statistics
almost periodic time-varying MA system
normal equation
identification of system
