UAV Target Tracking Technology Based on Improved GM-PHD Filter

Abstract

The integration of 6G technology and blockchain has promoted the booming development of UAV industry, but the accompanying national social security issues have also attracted public attention. In view of UAV's characteristics of “low, slow and small” and flexible turning, there is an urgent need for a technology that can eliminate background clutter and complete target tracking of multiple UAVs. At present, most of the existing researches tend to track the technology processing which depends on the detection results, and the detection performance is poor and the tracking instability is easy to occur. Based on GM-PHD filter and combined with machine learning, this paper proposes a DGM-PHD filter with high tracking performance, and evaluates it through MATLAB simulation software, using OSPA distance as evaluation index. The results show that the DGM-PHD filter performs better than the GM-PHD filter in the multi-UAV tracking problem under the uniformly accelerated linear motion, and achieves good tracking effect.