Distributed sampled-data state estimation for sensor networks with nonuniform samplings

Abstract

This paper is concerned with the distributed sampled-data H ∞ state estimation problem for a class of discrete time-invariant systems over sensor networks. The plant under consideration is sampled with a fast period. The sampling intervals of the measurements are integer multiples of the fast period and the sampling processing is characterized by a Markov chain. In order to estimate the plant state, a set of distributed estimators is proposed based on the randomly sampled measurements received by each sensor. The measurements received by each sensor include the information not only from the plant but also from its neighbors. By taking advantage of a Lyapunov functional approach, we first derive a sufficient condition under which the estimation error dynamics is stochastically stable and the H ∞ performance constraint is satisfied. Then, the desired distributed estimator gains are obtained by solving some matrix inequalities. In the end, the usefulness of the proposed estimation algorithm is verified by a numerical simulation example.