Abstract: How to consecutively track the boost phase and coast phase transitions without any prior information has become a pressing issue. Solving the problem with interacting multiple model is a feasible method. Since the Markov parameter of conventional interacting multiple model is constant, the model switching slows down. An adaptive Markov parameter IMM (AMP-IMM) algorithm is developed based on the modification of posterior information. The Markov transition probabilities can be modified adaptively in the process of filtering by omitting the information of non-matching model and magnifying the matching model information simultaneously during switching. The algorithm accelerates system convergence speed. Simulation results show that the proposed algorithm can consecutively track BM from the boost phase to the coast phase and is more effective than the conventional interacting multiple model.

Key words: ordnance science and technology, ballistic missile, extended Kalman filter, interacting multiple model, Markov matrix, posterior information