A Multiple Model Particle Filter for Maneuvering Target Tracking Based on Composite Sampling

A novel multiple model particle filter based on composite sampling is presented, which can decrease the number of particles required in the multiple model particle filter. It is proved that the algorithm based on composite sampling is a kind of multiple model particle filter. The computational complexity of the algorithm is similar to that of single model particle filter. Simulation shows that compared with the existing multiple model particle filter, the computational complexity of the proposed filter is decreased greatly.