Consensus-based distributed detection and tracking in clutter with random existence of target

We consider the distributed detection and tracking in the wireless sensor networks (WSNs) where the target may appear in or disappear from the surveillance region at random instances. We derive two consensus-based distributed detection and tracking algorithms. The first algorithm is for the distributed detection and tracking without clutter. With the cubature rule, we derive a centralized implementation with a fictitious fusion center (FC) and then a consensus-based distributed implementation without FC. The second algorithm is for the distributed detection and tracking in the clutter environment. Because of the clutter, it gives rise to more challenging scenarios. Based on the probabilistic data association filter, we develop a distributed algorithm for all nodes to cooperatively deal with the data association uncertainty. Numerical results illustrate that the distributed estimate at each node can be close to the global estimate with the FC which has the information of all nodes in the network.