The Trajectory PHD Filter for Jump Markov System Models and Its Gaussian Mixture Implementation

Based on the recently proposed set of trajectories, the trajectory probability hypothesis density (TPHD) filter is capable of producing trajectory estimates in first principle without adding labels or tags. In this paper, we propose a robust TPHD filter to track maneuvering targets through compatible the jump Markov system (JMS) model that the highly dynamic targets movement switches between multiple models, referred as JMS-TPHD filter. Firstly, we extend the concept of JMS to set of trajectories and derive the recursion for the proposed filter. Then, we develop the linear Gaussian Mixture (LGM) implementation of JMS-TPHD recursion and also consider the L-scan approximations of the implementations for computational efficiency. Finally, in a challgenging multiple maneuvering targets tracking scenario, the simulation results demonstrate the performance of the JMS-TPHD filter.