A Target Tracking Agorithm Based on EM Method with Inaccurate Based Stations

Aiming at the difficulty of target tracking caused by inaccurate base stations and the performance degradation of target tracking algorithms, a target tracking algorithm based on Expectation-Maximization(EM) method is proposed. Considering the inaccuracy of the base station locations, the algorithm sets the target state as a hidden variable and the base station location as an adjustable parameter, then jointly estimates them by EM method. In E-step, the joint posterior probability of the target under each base station is established, and the target state under each base station is estimated by particle filter. In M -step, a joint likelihood function based on the target and base station is established and the position of the base station is estimated by maximizing the joint likelihood function. Finally, based on the minimum mean square error criterion, the estimation results of each base station are fused to obtain the final solution of the target. The simulation results show that the algorithm shows excellent tracking performance and convergence in the environment of inaccurate base station locations.