Adaptive model set interacting multiple model auxiliary particle filter algorithm

To improve tracking accuracy, an adaptive model set interacting multiple model auxiliary particle filter (AMSIMMAPF) algorithm is presented, based on the turning model. The instant angular velocity is identified by the turning model and the model set is updated by this angle. The auxiliary particle filter can avoid weight degeneracy, sample the impoverishment and not restricted by assumptions of linearity or Gaussian noise. So it is selected in the respective model to improve its accuracy. The results of theoretical analysis and simulation verify that the algorithm has more precise filter result and less computation load, when compared with the interacting multiple model particle filter (IMMPF) algorithm.