H∞ filtering of a class of multi-rate multi-sensor fusion systems

This paper presents the problem of multi-rate filtering for a class of networked multi-sensor fusion systems: the interested state evolves according to a linear discrete-time model with unknown inputs (UIs); sensors are distributively deployed and their sampling rates are not identical; multi-rate sensor measurements, corrupted by UIs, are subject to stochastic packet dropouts (PDs) in the transmission to a fusion center for state estimation. Through transforming such a multi-rate filtering problem into design of a single-rate unknown input observer with periodical gains under a causality constraint, we turn to design an H_∞ filter, whose parameters are determined off-line via an iterative optimization of a set of linear matrix inequalities (LMIs). A numerical example of distributive multisensor target tracking is given to illustrate the proposed filter design.