Low-Altitude Multi-Target Tracking Technology Based On Twin LIDAR And IR Sensors

As the modern field environment becomes more and more sophisticated,the increase of t he physical noise disturbance in the low-airspace,the traditional single radar sensor detection and tracking mode of work is vulnerable to the impact of the physical noise in the airspace,Ha itian noise and weather environment.At the same time,because the detector has a certain elev ation angle limit,resulting in the lower limit of the elevation angle of the target can not be dete cted,can not accurately obtain the target’s spatial location information,and ultimately lead to t he target can not achieve accurate tracking.With the high speed growth in the field of sensor t echnology,the implementation of target tracking technology is also rapidly evolving.Modern battlefield tracking requirements have spanned from single target to multi-target tracking,and single sensor tracking is also developing to multi-sensor fusion tracking.For the urgent need o f multi-target tracking technology in complex low-altitude environment,this paper uses LIDA R/IR double sensor fusion ideology and digital information fusion filtering algorithm to impro ve the tracking precision of low-altitude multi-targetsA low-altitude single-target fusion tracking algorithm based on infrared sensors and LID AR is proposed based on the study of multi-sensor target tracking algorithms.The idea of sequ ential filtering tracking is combined with the LIDAR and IR fusion system to give full play to the detection advantages of the LIDAR and IR sensor fusion system.The interactive multi-mo del traceless Kalman filtering algorithm is added to the filtering process.By comparing the tra cking effect with that of radar,the multi-sensor fusion tracking algorithm shows more accurate tracking for a single maneuvering target in low altitude.For the problem that the target tracking in realistic low-altitude airspace is multi-target an d the motion form is also mostly nonlinear,the tracking algorithm of single radar sensor for no nlinear multi-target with different clutter intensity is studied.The multi-target probabilistic hy pothesis density filtering method is given,and the volumetric Kalman filter（CKF）is derived t o improve the GM-PHD filtering based on the traditional UKF improved GM-PHD filtering m ethod by establishing a nonlinear GM-PHD multi-target tracking model.Based on the problem of unrestricted growth of the number of Gaussian terms in the CKF-GM-PHD filtering algorit hm,an ACKF-GM-PHD filtering algorithm based on adaptive thresholds is given.The reliabil ity of the ACKF-GM-PHD filtering algorithm is confirmed by simulations,and the improved a lgorithm given in this paper is compared with the UKF-GM-PHD filtering algorithm in terms of target number prediction and target position state prediction.The simulation experiments sh ow that both algorithms can make reasonable estimates of target position state estimation and t arget number estimation in the low clutter background,but the ACKF-GM-PHD filtering can c orrect the estimates of target number and target position more quickly;in the dense clutter bac kground,both algorithms show some deviations in the target position state estimation and targ et number estimation,but overall the ACKF-GM-PHD filtering algorithm The filtering precisi on of the ACKF-GM-PHD filtering algorithm is generally better than that of the traditional U KF-GM-PHD filtering.To address the problem that a single detection regime cannot accurately track multiple tar gets at low altitude in a dense clutter background,a multi-target tracking algorithm based on A CKF-GM-PHD with LIDAR/IR sequential fusion is proposed based on the study of multi-targ et tracking algorithms.For the problem that the ACKF-GM-PHD filtering algorithm has poor t racking effect in dense clutter background,an adaptive gray wolf algorithm（AGWO）based o n ACKF-GM-PHD filtering algorithm is proposed to improve the ACKF-GM-PHD filtering al gorithm.The reliability and effectiveness of the AGWO-ACKF-GM-PHD algorithm are verifi ed by computer,and the improved filter fusion system based on the adaptive gray wolf algorit hm can significantly improve the accuracy of target number estimation and target position stat e estimation when compared with single radar tracking system and two-sensor fusion tracking system.