Research On Track Generation Problem In Multi-Target Tracking

Target tracking and track generation,as a major application of information fusion technology,have great practical significance in both military field and civilian applications.Especially in the military field,timely and accurate tracking of the state of enemy targets will greatly help us to do the following command.The problem of track generation involves many research contents.Based on multi-target tracking,this paper focused on maneuvering target tracking,multi-target fuzzy track correlation theory,and the structural design of distributed track generation system.Firstly,this paper introduced the basic theory of information fusion model.Because of the great advantages in reliability and accuracy,the distributed information fusion system is the basis of the research in this paper.Then,for the problem of spatial and temporal registration in multi-sensors multi-target system,several common spatial coordinate systems and coordinate transformation methods are introduced,and the temporal registration algorithm is studied.Secondly,for the target state estimation problem,the standard Kalman filter algorithm(KF)and two nonlinear filtering algorithms,extended Kalman filter(EKF)and unscented Kalman filter(UKF),are studied.Then aiming at the tracking problem of maneuvering targets,several common target motion models are introduced to match the different real motion states of the target.Then the interactive multi-model algorithm(IMM)is mainly studied,and the tracking performance of maneuvering target is analyzed by using IMM algorithm.Thirdly,the track correlation algorithms are studied for multi-target tracking problem.Because of the fuzziness in track correlation,the fuzzy theory is introduced and the fuzzy track correlation algorithms are mainly studied.The fuzzy C-means clustering(FCM)algorithm has a good effect on multi-target track correlation,but this algorithm only considers the target state at the current time.The correlation effect will decrease when there are more track crossings or bifurcations.In this paper,the relationship between the current track state and the historical state of the moving target is considered.And the idea of comprehensive fuzzy similarity is introduced.Combining the current track correlation calculation with historical correlation information,the improved FCM track correlation algorithm based on comprehensive fuzzy similarity is proposed.The simulation verifies that the algorithm has higher correlation accuracy than the classical FCM algorithm and statistical correlation algorithm.Finally,several common track fusion algorithms and the asynchronous track fusion algorithm for asynchronous data are introduced.In this paper,one multi-platform multi-sensor tracking system is also proposed.By designing the algorithm flow and the simulation environment,the simulation of the track generation system is carried out.And the simulation results show that in multi-target multi-sensor target tracking system,the tracking effect of the proposed multi-platform system is better than that of the single platform.