Distributed robust filtering with H∞ consensus of estimates

The paper addresses a problem of design of distributed robust filters using the recent vector dissipativity theory. The main result is a sufficient condition which guarantees a suboptimal H_∞ level of disagreement of estimates in a network of filters. The condition is formulated in terms of feasibility of certain linear matrix inequalities (LMIs). It provides a tractable algorithm for the design of interconnection gains for a network of distributed robust observers. The proposed approach is applied to the problem of observer-based robust synchronization of a nonlinear network to an isolated node.