基于辅助模型的递推增广最小二乘辨识方法

Recursive extended least squares identification method based on auxiliary models

中文关键词  递推辨识  参数估计  最小二乘  辅助模型  输出误差滑动平均模型

英文关键词  recursive identification  parameter estimation  least squares  auxiliary models  output error moving average models (OEMA)

中文摘要

      针对有色噪声干扰的输出误差滑动平均系统, 将辅助模型与递推增广最小二乘算法相结合: 用辅助模型的输出代替辨识模型信息向量中的未知真实输出项, 用估计残差代替信息向量中的不可测噪声项, 从而提出了基于辅助模型的递推增广最小二乘辨识方法. 为了展示所提方法的特点, 文中还给出了经过模型变换的递推增广最小二乘算法. 理论分析和仿真研究表明, 提出的方法原理简单、计算量小, 可以给出高精度参数估计, 且能够用于在线辨识.

英文摘要

      For output error moving average systems with colored noises (OEMA model), this paper combines the auxiliary model and the recursive extended least squares algorithm to present the auxiliary model based recursive extended least squares (AMRELS) algorithm. In this approach, we replace the unknown true outputs in the information vector with the outputs of the auxiliary model, and replace the immeasurable noise terms in the information vector with the estimated residuals. To demonstrate the advantage of the proposed algorithm, this paper gives the recursive extended least squares algorithm through model transformation. The analysis and simulation results indicate that the AMRELS algorithm is simple in principle, with less computational burden, capable of generating more accurate parameter estimates and can be implemented on-line.