多维ARMAX系统辨识的一种新的递推预报误差算法
A NEW RECURSIVE PREDICTION ERROR ALGORITHM FOR IDENTIFICATION OF MIMO ARMAX SYSTEMS
-
摘要: 本文根据预报误差（RPE）算法的思想，对多输入-多输出ARMAX模型A（z）yn=B（z）un+C（z）εn发展了一种新的辨识算法．由于在算法中加了矩阵多项式C（z）的稳定性检验，这一种算法比推广最小二乘算法具有更快的收敛速度，这个结果在计算机仿真中得到了证实．Abstract: Based on the theory of recursive prediction error method,a new identification algorithm is developedfor the multiinput multi-output ARMAX model A(z)yn=B(z)un+C(z)ε.The stability of polynomial matrix C(z) is tested in this algrithm.Simulation of an actual system shows that the convergent rate of thisalgorithm is greater than that of the extented least square(ELS)algorithm.
下载:
