一种辨识多变量系统结构和参数地递推算法*

本文基于Guidorzi方法,提出一种辨识多变量系统结构和参数的递推算法。本算法通过对矩阵进行分解,根据分解过程中某参数的变化情况,判断矩阵的奇异性,从而辨识出系统的结构和参数,使得原来对矩阵求逆及求行列式的运算变为简单的代数运算,大大减少了辨识过程中的计算量,数值实例表明了这种算法的有效性。

A Recursive Algorithm for the Identification of Structure and Parameter of Multivariable Systems

In this papcr,a recursive identification algorithm for determining structural indices and parameters of multivariable system is presented on the basis of Guidorizi's method. This algorithm is based on matrix decomposition by which the singularity of matric is determined according to the variation of a certain parameter is decomposition, therefore the structure indices and parameters of system arc estimatcd.In the whole algorithm it is necessary ncither to evaluate the value of determinant nor to inverse the matrix ,requiring only simple algebra operations. So,the amount of calculation is identification is greatly reduced . The validity of proposed algorithm has been proved by the results of digital simulation.