摘要
把有色噪声干扰的随机系统看作一类广义输出误差模型,辨识的思想是把过程模型和噪声模型分别用两个有限脉冲响应(FIR)模型逼近,进而得到一个特殊的CARMA模型/ARMAX模型。使用增广最小二乘法估计其参数,最后根据模型等价原理确定原系统的参数估计。仿真结果表明．如果近似的FIR模型有较高的阶次,系统辨识精度满足要求。
Systems with colored noises are regarded as a class of generalized output error systems. The basic idea is to approximate the process model and the noise model by using （FIR） models, thus obtaining a special CARMA/ARMAX model. The model can be identified by the extended least squares. Finally, original system parameters are determined by the model equivalence principle. Simulation results indicate that if FIR models have high orders, the system identification satisfies the precision demands.
出处
《南京航空航天大学学报》
EI
CAS
CSCD
北大核心
2006年第B07期167-171,共5页
Journal of Nanjing University of Aeronautics & Astronautics
基金
国家自然科学基金(60574051)资助项目。
关键词
输出误差模型
参数估计
辨识
最小二乘
output error models
parameter estimation
identification
least squares
