Distributed optimal fusion estimation for multi-sensor systems subject to random delay and packet drop

This paper considers the distributed optimal (i.e., linear minimum variance) fusion estimation problems for two classes of networked multi-sensor systems. In the first class of systems, the sensors´ measurements are transmitted via unreliable digital communication networks (DCN) to local estimators for obtaining local estimates of the state. Then the local estimates are fused in the fusion center to get the fusion estimate. The data transmission from local estimators to the fusion center is not via DCN. In the second class of systems, the sensors´ measurements are processed locally to obtain local estimates of the state. Then, via DCN, the local estimates are transmitted to the fusion center where they are fused to get the fusion estimate. In these systems, the data transmission via DCN is subject to random delay and packet drop. The results of local estimation have been presented in the literature. The estimation error cross-covariances between local estimates are derived in this paper. By the fusion rule weighted by matrices, the distributed optimal fusion estimators are developed.