一类系统参数辨识的按段多重一般正交多项式方法

PIECEWISE MULTIPLE GENERAL ORTHOGONAL POLYNOMIALS METHOD TO PARAMETER IDENTIFICATION FOR A CLASS OF SYSTEM

本文介绍了按段多重一般正交多项式系及其基本性质,并把它们应用于参数可分离系统的参数辨识.由于采用了按段低阶正交多项式多重逼近技术,该方法具有计算量少、结果精度高、可递推计算及不需要被辨识参数的初始估计等优点.本文提出了两个算法,并成功地应用于发酵过程细菌生长动力学模型的参数辨识.

This paper introduces a serics of piecewise multiple general orthogonal polynomials,presents anddemonstrates its basic properties and applies them to parameter identification for parameter separable sys- tems.Due to the use of approximation technique by piecewise multiple low order orthogonal polynomials,the method manifests the following advantages:the amount of computation is little,the accuracy of resultsis high,it can be recurrently calculated and the algorithms do not need the initial estimation of the esti-mated parameters.Two algorithms have been presented and successfully applied to parameter identifica-tion for bacteria growth dynamic model of fermentation process.