A probability hypothesis density filter with Singer model for maneuver target tracking

摘要

With the purpose to solve the target loss problem of PHD filter in maneuvering targets tracking, new methods that combines the Singer model with mixture Gaussian (Singer-GMPHD) filter is proposed. This method is based on mixture Gaussian probability hypothesis density filter. modeling the Gaussian components with Singer model. Then the Gaussian componentsare updated with traditional PHD filter. Simulation results indicate that this method gives perfect performance on tracking maneuvering targets movement with unknown targets number by combine the features of both PHD filter and the Singer model. And the accuracy of estimation of targets number is improved. This method shows the number of targets estimated by the proposed algorithm is consistent with the real situation. And the OSPA distance value that describes the estimation error decrease evidently.