Sequential Inverse Covariance Intersection Fusion Kalman Filter for Multi-sensor Systems with Packet Dropouts

This paper mainly studies the fusion estimation problem of multi-sensor systems with packet dropouts. Firstly, by the augmented state algorithm, the state space model is transformed into a new augmented system with fictitious noises. By Kalman filtering method and Inverse Covariance Intersection (ICI) method, Sequential Inverse Covariance Intersection (SICI) structure is designed and then the SICI fusion Kalman filter is presented, which avoids the computation of the cross-covariances among the local filters and has less conservativeness than the previous Sequential Covariance Intersection (SCI) filters. The accuracy relations among the local estimators and the fused estimators are proved. The validity and the consistency of the presented fuser are demonstrated by a simulation example, and its precision relation is shown.