Consensus-based Distributed Kalman-Bucy Filter for Continuous-time Systems

Abstract

In this paper we introduce a distributed consensus-based Kalman filter for distributed state estimation of continuous-time systems. In particular, we achieve stability of the estimation error only assuming that all agents together are able to observe the system, which is in contrast to each agent possessing this property individually. The algorithm is implementable without any precomputation of filter parameters, such as coupling strength, or global knowledge about the graph.