摘要
本文对连续-离散系统的极大似然估计方法提出了一种灵敏度初值的正交计算法,它是对具有递推计算灵敏度的修正牛顿-拉夫森算法的改进。文中通过借用正交试验的概念和正交表的性质,解决了原有算法的初始矩阵求逆所存在的困难。
This paper presents an orthogonal computation method of initial value of sensitivity for maximum likelihood estimation method to he continuous-discrete system. This method is an improvement to the modified Newton-Raphson scheme with estimated sensitivities (MNRES) method. The difficulty to invert the initial matrix in MNRES method is solved by the concept of orthogonal test and orthogonal table
出处
《自动化学报》
EI
CSCD
北大核心
1992年第5期619-622,共4页
Acta Automatica Sinica
基金
国家自然科学基金资助项目
关键词
算法
极大似然
灵敏度
Continuous-discrete system
maximum likelihood estimation
sensitivity
orthogonal test.
