An efficient algorithm unifying MHT and Multiple Model approaches

Main principles for multiple target tracking are data association and state estimation. Multiple Hypothesis Tracking (MHT) and Multiple Models (MM) methods have been the prevailing data association and state estimation methods for maneuvering targets tracking. A MM/MHT algorithm which incorporates multiple model mechanics into MHT framework has been designed and implemented in this paper. The algorithm is tested for a typical scenario in simulation, the results show that MM/MHT algorithm is effective and feasible to use in real tracking systems.