Abstract:

For the problem of multisensormultitarget tracking in nonlinear systems, a modified parallel implementation of multisensor unscented multiple hypothesis tracking algorithm (MPUMHT) is proposed. In the new algorithm, in order to solve the problem of error cumulation of the parallel multisensor algorithm, the estimation from each sensor is weighted according to the method of probabilistic data association first and a modified parallel multisensor algorithm is derived. Then a multiple hypothesis tracking method is used for the association of measurements to tracks. Finally, the problem of target tracking in nonlinear systems is implemented according to unscented Kalman filter (UKF), and the MPUMHT algorithm is derived. According to the simulation results, the accuracy and robustness of the proposed algorithm are improved compared with the multisensor joint probabilistic data association/extend Kalman filter (MSJPDA/EKF) algorithm.