Asynchronous networked estimation system for continuous time stochastic processes

In this paper we examine an asynchronous networked estimation system for state estimation of continuous time stochastic processes. Such a system is comprised of several estimation nodes connected using a possibly incomplete communication graph. Each of the nodes uses a Kalman filter algorithm and data from a local sensor to compute local state estimates of the process under observation. It also performs data fusion of local estimates and data received from other directly connected nodes. Asynchronism means that each node preforms measurement, estimation, and data fusion in moments of time that are independent of the work cycles of the other nodes.