Local node cooperative estimation of stochastic systems using a switching network of heterogeneous sensors with communication delay

This paper addresses the cooperative stochastic state estimation of a moving target using a switching sensor network that may contain communication delays. The switching in the graph of the sensor network may be due to local communication failures of the neighboring sensors or due to their movements. The network may have heterogeneous sensors with different measurement vectors of various combinations of states. Using a Kalman filter, the states of the target system are estimated at each node having only its local neighboring information. The network nodes cooperate locally with a packet-based communication to improve the estimation. Finally, numerical examples are provided to show the effectiveness of the scheme. The results are compared with the centralized and local estimations.