Distributed observer for linear systems with bounded unknown inputs

Abstract

In this paper, a fully distributed unknown input observer is designed, which can give an exact state estimate for linear time-invariant systems with unknown inputs. Specifically, it is a group of sensors with complementary measurement capabilities, in which each one serves as a local observer and has communication links with neighbors. Such networked sensor nodes are designed to cooperate by taking complementary actions, so that each one can estimate the full state of the plant exactly. The fully distributed design in this paper requires no global information such as the communication topology, collective measurement capability, and bound of the unknown input.