Steady State Filters Bank Based Interacting Multiple-Model-Extended Viterbi Algorithm for Maneuvering Target Tracking

Abstract

The performance of Interacting Multiple-Model (IMM) algorithm depends on model set and filters bank. Most of the improved IMM algorithms originate from designing more reasonable and effective model set structures. One example in case is the Interacting Multiple-Model-Extended Viterbi (IMMEV) algorithm, which can not only inherit the effective cooperation strategies of the IMM algorithm, but also adapt to the outside world like variable structure IMM algorithm by eliminating those mode transition probabilities that are harmful. However, this design approach is difficult to achieve significant tracking performance and substantial complexity reduction, simultaneously. In this paper, the steady state filters bank based Interacting Multiple-Model-Extended Viterbi algorithm for maneuvering target tracking is presented. The main idea of the algorithm is to further improve the IMM-EV algorithm performance and reduce its complexity by designing a new filters bank. The new approach is based on the steady state filters (ab andabg filters) bank using IMM-EV algorithm design instead of a Kalman filters (second and third order filters) bank. Real maneuvering target tracking application experiments demonstrate that the SS-IMM–EV algorithm with some suitable choice of ”m” is superior to the IMM, Fast-IMM and IMM-EV algorithms, which can obtain a considerable tracking performance and reduce complexity markedly.