Consensus-Based Distributed Multitarget Tracking with Probabilistic Track-to-Track Association

In this paper a novel consensus algorithm for distributed multi-sensor multi-target tracking is proposed, adaptive in the context of sensors having limited sensing range in a cluttered environment. The algorithm is composed of two consensus schemes working in parallel; it assumes intersensor communication of target state estimates, together with adaptive scalars used for weighting the estimates. A new trackto-track association and fusion methodology, matching the communicated estimates of different sensors, is also proposed. It is based on the all neighbors approach and the Integrated Probabilistic Data Association methodology. An accompanying theoretical analysis of stability and reduction of noise influence gives an insight into the interplay between the local trackers and the global algorithm. Numerical simulations illustrate specific properties of the algorithm and show that it exhibits very good tracking performance, close to the centralized solution.