Multi-sensor Box Particle Filter with Iterated Measurement Contraction

Ahstract-The box particle filter is a time-saving alternative to the particle filter for nonlinear multi-target tracking in the presence of interval measurements. Box particle implementations of single-sensor multi-target tracking algorithms have been developed in the literature, which involve the contraction of box particles under the constraints imposed by the set of measurements from the sensor. However, tracking with multiple sensors all returning interval measurements still remain an unanswered issue for box particle filtering. Namely, there is no available mechanism for the contraction of box particles given multiple sets of measurements from multiple sensors. In this paper, a novel mechanism is proposed that features iterated contraction of box particles under the constraints imposed by the sets of measurements from the sensors. Based on the proposed mechanism a multi-sensor box particle implementation is developed. Simulation results show that the proposed mechanism is effective for the considered issue.