Tracking of Underwater Maneuvering Target via M-SIMMUKF Algorithm

For localizing and tracking of an underwater maneuvering target by sonobuoys, a novel algorithm is proposed based on the scalar weight interactive multiple model (SIMM) and the unscented Kalman filter (UKF). Firstly, the SIMMUKF algorithm is directly obtained by combining the SIMM with the UKF. However, the filter estimation errors corresponding to each model are assumed to be uncorrelated in the SIMMUKF algorithm. In order to solve this problem, the multi-sensor optimal information fusion under the criterion of linear minimum variance is utilized to correct the estimated probability in the SIMMUKF algorithm. Then, a target motion model probability correction algorithm referred to as Modified-SIMMUKF (M-SIMMUKF) is obtained. The tracking accuracy of the M-SIMMUKF algorithm is higher. Numerical results validated the proposed scheme.