Sequential fusion algorithm based on unscented particle probability hypothesis density filter

In the case of clutter,missed detections and no-linear,the single sensor particle probability hypothesis density filter(P-PHDF) algorithm will result in many problems,such as low accurate,filter divergence and particle degradation.To overcome these problems,an unscented PHD filter(UP-PHDF) for multi-sensor multi-target tracking based on sequential fusion is proposed.Firstly,the unscented particle filter(UPF) is employed to fulfill PHDF,and the unscented Kalman filter(UKF) method is applied to generate and sample the importance density function.The method could make the distribution of particles much close to the distribution of multi-target PHD.Secondly,in order to improve the accuracy of the algorithm,based on radar and infrared sensor to achieve the sequential fusion algorithm of the UP-PHDF,the two sensors alternately filtering is designed to guarantee observability of the target state.Simulation results demonstrate that the accuracy and stability of proposed algorithm for tracking systems are greatly superior to the single P-PHDF in complicated case.