Research On Asynchronous Track Fusion Algorithm For Multisensor Maneuvering Target Tracking

Modern warfare environment is complex with various forms of information.On the battlefield valuable targets are monitored all the time by vast number of detection equipment.In multi-sensor detection system,sensors work in an independent way and sample in different frequency,which brought the problem of asynchronous measurement fusion.In order to develop an asynchronous target tracking system,a pseudomeasurement based asynchronous fusion algorithm is researched both in centralized and distributed fusion structure.Then the problem of maneuvering target tracking is discussed.An asynchronous track fusion algorithm with nonlinear motion model is proposed based on interactive multiple model algorithm.Furthermore,as the model factor in the outputs of pseudo-measurement based fusion algorithm is over-weighted,a secondary filter method is used to improve the asynchronous track fusion algorithm for maneuvering target tracking.With the progress in research,the influence of structure and frequency of the proposed fusion algorithm on tracking performance is studied,the asynchronous track fusion algorithm with information feedback and with sequential method are separately presented.The specific contents are as follows:(1)The pseudo-measurement based asynchronous fusion algorithm is first researched in centralized tracking system.In distributed tracking structure,information filter is widely used,and is thus compared with Kalman filter,the result of which shows the equivalence.Then the pseudo-measurement based asynchronous track fusion algorithm is presented,and the simulation results show that fusion algorithm has better performance with lower sampling frequency.(2)Interacting multi-model algorithm(IMM)is a typical maneuvering target tracking method.The pseudo-measurement based asynchronous fusion algorithm uses the linear motion model,which is rarely practical.By applying asynchronous track fusion method into the parallel filtering section of the IMM,an asynchronous track fusion algorithm for maneuvering target tracking is proposed.Moreover,a secondary filter method is brought to calculate the model output and probability.(3)The influence of structure and frequency of the proposed fusion algorithm on tracking performance is discussed.The influence of fusion structure to the output of information matrix fusion(IMF)is studied,which is a special form of the pseudomeasurement based asynchronous fusion algorithm.Then the conclusion is applied to the asynchronous fusion case,the simulations of which show the improvement of local tracks.(4)When a fusion center with strong calculating ability is used,the pseudo-measurement based fusion algorithms does not make full use of short fusion interval.In the thesis,a sequential method based asynchronous track fusion algorithm for maneuvering target tracking is proposed with simulations showing its advantages in tracking performance.