A new multi-target state estimation algorithm for PHD particle filter

Abstract

Probability hypothesis density (PHD) filter is a new practical method to solve the unknown time-varying multi-target tracking problem. Particle filter implementation of the PHD filter has demonstrated a feasible suboptimal method for tracking multi-target in real-time. To obtain the target states, the peak-extraction from the posterior PHD particles needs to be implemented. A new state estimation method is proposed in this paper, which doesn't need to extract the PHD peaks. The method provides a single-target PHD expression derived from the updated PHD equation. The single-target PHD is approximated by the particles and their weights relevant to the observation. Thus the target states can be directly estimated from the single-target PHD sequentially. Simulation results demonstrate that the new algorithm provides more accurate state estimations and is more efficient than the traditional multi-target state estimation methods such as k-means clustering algorithm.