Multi-Platform Multi-Target Tracking And Data Association Algorithm Research

In recent years,multi-object tracking and data association techniques used in multi-object tracking have gradually become hot topics in academic research.Hot research projects in the field of underwater acoustics,such as multi-unmanned platform collaborative detection and tracking,and multi-base network detection and tracking,all require multi-object tracking and data association technology as support.Previous target tracking and data association techniques used in the field of underwater acoustics often made little use of information other than target motion information.This paper will mainly focus on the systematic study of key algorithms including data association for multi-platform multi-object tracking technology,and propose a data association algorithm based on target acoustic comprehensive features.For the multi-platform multi-target tracking model,this paper studied the existing methods for modeling the motion model of underwater targets,passive pure bearing positioning algorithms,extended Kalman filter methods and particle filter methods suitable for nonlinear models.Classic data association algorithms were analyzed and deduced.The root mean square error of the position and the generalized optimal sub pattern assignment index were studied and used as performance evaluation indicators for tracking algorithms.In the context of multi-platform multi-object tracking,the problem of generating a large number of false localization points with pure bearing localization was addressed by using the generalized S-distribution(S-D)assignment algorithm and pure bearing passive localization algorithm to convert the problem into a static allocation problem between measurement results and localization target results.After calculating the localization results,the joint probability data association algorithm or the comprehensive joint probability data association algorithm was used to dynamically associate the localization target points and tracks.Through simulation,the tracking efficiency and accuracy of different algorithms in the multi-platform multi-object tracking scenario were verified.The signals received by each measurement unit in the underwater passive localization and tracking system often contain a large amount of physical information about the target.This paper proposes a data association method based on target acoustic comprehensive features.The characteristic extraction technology of underwater target radiation noise was studied,and a multidimensional comprehensive feature for underwater target radiation noise was constructed and integrated into the data association algorithm.By using the target’s own acoustic physical information,the accuracy of association is improved,and the formation of false measurement trajectories is suppressed.Through simulation verification and performance comparison of different algorithms,the accuracy and reliability of the proposed algorithm were verified.Finally,the improved algorithm was used to process the data obtained from the lake experiment to verify the effectiveness of the algorithm in processing measured data.