论文部分内容阅读
本文提出了一种称为模型群切换算法(MGS)的通用的变结构多模型(VSMM)估计器.它假设整个模型集合可以用一定数目的模型群来覆盖,每个模型群代表相互间紧密联系的一族系统行为方式或结构,在任何给定时刻,由一个硬决策来决定具体运行的模型群。它是第一种可普遍应用于一大类具有混合(连续或离散)不确定性问题的VSMM估计器。同时,它很容易实现。通过一个简单的失效检测和识别的例子显示当具有与固定结构的交互多模型(FSIMM)估计器相同的性能时,MGS算法可以从本质上减少计算量。
In this paper, we present a general variable structure multi-model (VSMM) estimator called Model Group Switching Algorithm (MGS), which assumes that the entire model set can be covered by a certain number of model groups, each representing a close The family of linked systems behaves in a manner or structure that, at any given moment, is determined by a hard decision to the model group that is operating specifically. It is the first VSMM estimator that can be universally applied to a broad class of problems with mixed (continuous or discrete) uncertainties. At the same time, it is easy to implement. A simple example of failure detection and identification shows that the MGS algorithm can substantially reduce the computational load when it has the same performance as the FSIMM estimator with fixed structure.