Study On Theory And Algorithm For Maneuvering Target Tracking

The maneuvering target tracking is a research exercise to the tracking target which can't accurate description of target motion estimation. Therefore, multi-target tracking is the processing of measurements that is received by sensors to maintain the state estimates of multiple targets. Such technology is one of the most important topics in national defense research such as radar, sonar signal processing, and in the military and civilian. It has a wide application. In this dissertation, the basic theories and methods of the tracking target are introduced; Based on the continuous Hopfield neural networks and simulated annealing algorithm combined to overcome the joint probability data association algorithm in the composition of the explosion, the evolution of network convergence slow and easy local minimum and other issues, which is discussed; The adaptive filtering parameters of the target tracking method is researched; Research is also based on the extended kalman particle filter of the maneuvering multi-target tracking technology, and in this regard, it has done some analysis and improved work. The main contributions are summarized as follows:1.A basic algorithm(EKF) is researched. About the maneuvering multi-target tracking technology, JPDA in the problem of the large amount of calculation, is based on continuous Hopfield neural network, to improve its processing speed, which combines the simulated annealing algorithm, in order to address the evolution of network convergence slow and easy local minimum and other issues. The simulation results are given and it makes the corresponding analysis of the network parameters.2.Concerning adaptive filter for the target tracking technology, this dissertation proposes a new adaptive filtering method, which mainly uses genetic algorithm to optimize filtering parameters in the maneuvering multi-target tracking. After the fitness function through the identification, genetic iteration through the current operation can find the corresponding filtering parameters with the premise of the best tracking. The simulation results show that the method improved the accuracy of the multi-target tracking.3.The elementary particle filter is reviewed. The multi-target tracking algorithm which is based on the mixed filter, including the extended kalman filter and the particle filter, is research. This algorithm is improved and the tracking result of the new algorithm is compared with the original one's. The simulation results confirm that the improved algorithm for the multi-target tracking increases to a certain extent of the overall accuracy, and also show that the superiority of the new algorithm.Finally, a comprehensive summary, concluding remarks and the further work are presented.