Maneuvering target tracking based on an adaptive variable structure interactive multiple model filtering and smoothing algorithm

For maneuvering target tracking, the interactive multiple model (IMM) algorithm and its variants have shown good performance, among which variable structure IMM (VSIMM) is most widely studied one. The approximation degree of matching subset will affect performance VSIMM. In addition, considering that smoothing can provide better estimation state, a new adaptive VSIMM filtering (AVSIMMFS) proposed in this paper. First, an accurate established, run parallel with independently from different subsets, state highest probability selected as final result. Then, (AVSIMM) was designed based on algorithm. adaptation reflected construction original improves between actual target. Finally, by data AVSIMM algorithm, AVSIMMFS obtained. Because combination forward backward smoothing, tracking accuracy further improved. Simulation results show than other methods.