Distributive target tracking in sensor networks with a markov random field model

Tracking in sensor networks has shown great potentials in many real world surveillance and emergency system. Due to the distributive nature and unpredictable topology structure of the randomly distributed sensor network, a good tracking algorithm must be able to aggregate large amounts of data from various unknown sources. In this paper, a distributive tracking algorithm is developed using a Markov random field (MRF) model to solve this problem. The Markov random field (MRF) utilizes probability distribution and conditional independency to identify the most relevant data from the less important data. The algorithm converts the randomly distributed network into a regularly distributed topology structure using cliques. This makes tracking in the randomly distributed network topology simple and more predictable. Simulation demonstrate that the algorithm performs well for various sensor field setting, and for various target sizes.