Generalized relaxation techniques for robust H∞ filtering of nonhomogeneous Markovian jump systems

摘要

This paper is concerned with the robust H∞ filtering problem for continuous-time nonhomogeneous Markovian jump systems via generalized relaxation techniques. To provide more general description on transition rates, this paper first presents four (or five) specific cases that can be classified according to the property of transition rates. And then, two generalized relaxation methods are proposed so that they can embrace all possible cases identified from our classification, even the excluded case from other results. Especially, to develop a less conservative relaxation process with reduced computational complexity, this paper proposes an improved approach that imposes the summation constraint of transition rates without resorting to extra slack variables, and takes advantage of their component-summation-based constraints.