H2H2 state estimation for network-based systems subject to probabilistic delays

In this paper, the estimation problem for network-based systems is investigated. The networked system is with limited communication capacity and the network-induced delay from the sampler to the estimator is considered. By studying the network characteristic, we assume that the probability of occurrence for each delay within a known set is known. The network-induced delay is detectable at the estimator node. We propose to design a class of delay-scheduling estimators. With the state augmentation technique, we obtain a stochastic estimation error system. The stochastic stability of the estimation error system is firstly studied. In order to evaluate and suppress the effect of the external input, the H2H2 performance is employed. Based on the derived sufficient conditions, we propose an iterative algorithm to compute the estimator parameters. A numerical example is given to illustrate the effectiveness of the proposed algorithm.

► Consider the network-induced delay which has a corresponding occurrence probability for possible delay.
► Design the H2H2 state estimator for network-based systems.
► Separate the Lyapunov matrix and the system matrix.
► Achieve less conservative results.