An improvement to the interacting multiple model (IMM) algorithm

Computing the optimal conditional mean state estimate for a jump Markov lin ear system requires exponential complexity, and hence, practical filtering algorithms are necessarily suboptimal. In the target tracking literature, s uboptimal multiple-model filtering algorithms, such as the interacting mult iple model (IMM) method and generalized pseudo-Bayesian (GPB) schemes, are widely used for state estimation of such systems. In this paper, we derive a reweighted interacting multiple model algorithm. Although the IMM algorit hm is an approximation of the conditional mean state estimator, our algorit hm is a recursive implementation of a maximum a posteriori (MAP) state sequ ence estimator. This MAP estimator is an instance of a recent version of th e EM algorithm known as the alternating expectation conditional maximizatio n (AECM) algorithm. Computer simulations indicate that the proposed reweigh ted IMM algorithm is a competitive alternative to the popular IMM algorithm and GPB methods.