The paper presents a distributed human tracking system based on pyroelectric sensors. The developing process of the system mainly includes two aspects: distributed tracking algorithm and self-calibration scheme. In the tracking algorithm, a distributed information filter is constructed with joint probabilistic data association (JPDA) and consensus method. The self-calibration utilizes Kullback-Leibler divergences to construct objective function and information projection to realize calibration process. A distributed message passing scheme is developed among neighboring sensor nodes to get distributed calibration. The simulation and experimental results verified the validity of tracking algorithm and improved tracking performance after using the proposed distributed self-calibration.
展开▼