The Research On Data Association In Multi-target Tracking

Data Association is the pivotal part of the target tracking system,as well as the problem in the field of multi-target tracking.With the development of the transport industry and the improvement of safety awareness,multi-target tracking technology is facing more and more severe challenges.As the core issue of tracking,the success of data association will directly affect the results of target tracking.Traditional data association algorithm which is likely to cause unsuccessful association,incorrect tracking and tracing missing is based on the kinematic parameters such as position and speed.In order to improve the associated performance and tracking accuracy,the paper mainly studies the data association method which is based on the amplitude information,size information and auxiliary information of AIS.Firstly,the paper analyzes the basic theory of multi-target tracking and data association including the target model,Discrete Kalman filter algorithmand common data association algorithms.The main focus of this part is the nearest neighbor,probabilistic data association and joint probabilistic data association algorithm.Simulations discuss the advantages and disadvantages of each method and provide a theoretical basis for the improved algorithm of the future.Secondly,through the analysis of the radar echo characteristics,this paper studies a kind of auxiliary correlation algorithm based on the amplitude and size information of target echoes,named PDA-AI and PDA-DI.In order to extend the applicability of the algorithm for different kinds of targets,the paper carries out an adaptive-weighted distribution factor according to the associated probability of the single feature,and fuses it into the multi-feature joint auxiliary association algorithm,named APDA-DI-AI.Through the value of correct association probability and filtering error curve,the effectiveness of the algorithm is verified respectively.Finally,the paper studies a kind of association algorithm based on the information of AIS.On the basis of the traditional track algorithm of grey correlation degree,it uses the theory of sequential indicator to improved correlation coefficient formula of absolute difference,and puts forward a track association algorithm which based on the sequential theory of sequential modified grey association degree to conduct track correlation for radar and AIS.The auxiliary correlation algorithm of AIS is realized using the track of each candidate point within radar tracking gate to establish a link between fusion weight quality factors.The simulation results show that the proposed algorithm is reliable and effective.