A Multi-directional Motion Interacting Fusion Model for Diver Tracking

Article history: Received: 10.04.2015. Received in revised form: 13.05.2015. Accepted: 29.05.2015. According to the diver motion characteristics, which are low speed and rapid change of direction, a multi-directional motion model is presented. Then the motion model is introduced into an interacting multiple model method, while the timevarying motion model transition probability was corrected according to current measurements. Firstly, the predictive state was obtained by a multi-directional motion model. Secondly, the parallel Kalman filters were applied to estimate multi-directional state. Finally, the interactive fusion processing for estimations from multidirectional motion model was conducted to implement diver state estimation. The method was verified by both simulation and experiment. The results show that the proposed method has higher tracking accuracy and superior adaptability than conventional interactive multiple model algorithm based on single direction motion model. The proposed method is effective for diver tracking.