Consensus based decentralized estimation for linear plants with unknown inputs

This paper addresses the problem of consensus-based decentralized state estimation for a class of linear time-invariant systems affected by stochastic disturbances and deterministic unknown inputs. All the local system models are assumed to fulfill the property of “strong detectability”. It is shown that a stable consensus based estimator exists, and an optimization procedure for computing the consensus gain parameters is described. A method that allows to reconstruct the unknown inputs acting on the system is also introduced. All the procedures and methodologies are verified by means of a simulation example.