On the convergence of distributed estimation of LTV dynamic system with switching directed topologies and time-varying sensing models
Shaocheng Wang, Wei Ren · 2016
In this paper, the problem of linear dynamic state estimation using distributed networked agents, is studied. Based on our previous work [1], the proposed algorithm is further extended to a more realistic scenario where both the process model of the target, and the local sensing model of each agent might be time varying. In addition, the directed communication topology, as adopted in [1], is further allowed to change with time. Moreover, the set of agents directly sensing the target is also subject to change over time in the extended scenario. The sufficient conditions for the approximated local estimate error covariances to be uniformly upper bounded in the positive definite sense, are formulated.