Robust filtering for uncertain networked systems with randomly delayed and lost measurements

This paper investigates the problem of robust filtering for networked linear time-varying systems with norm-bounded modelling uncertainties. Delay and dropout of measurement data in the transmission from sensor to filter are both modelled by a Bernoulli distributed random sequence. Finite-horizon two-stage Kalman filters are introduced whether or not the data packets in the network are time stamped. The parameters of robust Kalman filters are determined such that the covariances of the estimation errors are bounded and these upper bounds are guaranteed to be minimal. Novel augmented state vectors are used to extract procedures for computation of filters' parameters. Lastly, simulation results are presented to demonstrate the superior performance of the proposed approach compared the existing methods in the literature.