Analysis of Optimization Mechanism of the Hyper H_∞ Filter and Improvement of Its Fast Algorithm

点击翻译按钮获取中文摘要

The hyper H_∞ filter provided excellent convergence, tracking, and robust performances for system identification, but the mechanism of optimization in the hyper H_∞ filter has not been analyzed sufficiently. This study has clarified its optimization mechanism by solving a generalized hyper H_∞ filtering problem which further optimizes the H_∞ filter over a set of models and norm weights, providing a unified framework to various adaptive algorithms. Also, the error propagation of the fast algorithm has been analyzed, which leads to a numerical stabilization method.