Ground Moving Target Tracking with VS-IMM Using Mean Shift Unscented Particle Filter

摘要: In order to track ground moving target, a variable structure interacting multiple model (VS-IMM) using mean shift unscented particle filter (MS-UPF) is proposed in this paper. model-conditioned filtering, sample particles obtained from the are moved towards maximal posterior density estimation of target state through shift. On basis stop VS-IMM, hide proposed. Once obscured by terrain, prediction at prior time used instead measurement time; addition, road set not changed. A indication (GMTI) radar employed three common simulation scenarios target: entering or leaving road, crossing junction and no measurement. Two evaluation indexes, root square error (RMSE) average normalized squared (ANEES), used. The results indicate that when on which changes, tracking accuracy effectively improved algorithm. Moreover, interruption could be avoided if too slowly masked terrain.