Lane-Level and Full-Cycle Multivehicle Tracking Using Low-Channel Roadside LiDAR

The ability to track multiple objects is crucial for roadside units to provide high-precision, trajectory-based traffic data, especially for connected vehicles that require complementary and long-range traffic information to improve road safety. Occlusion and continuous tracking are major challenges that have hindered the achievement of obtaining accurate, uninterrupted, and consistent multiobject tracking using roadside light detection and range (LiDAR) technology. This article presents a lane-level and full-cycle multivehicle tracking (MVT) method that uses low-channel roadside LiDAR (LRLT). First, a lane-level map was created by analyzing multiple frames of traffic object detection results. Then, we introduced an association method based on a search process and a microscopic motion model, while considering the lane-level map as a constraint. The search process aims to identify all the previous tracks that might be linked to the target vehicle, while the microscopic motion model ensures finding a correct historical track for association. In addition, a detection optimization approach using the lane-level map was also proposed to enhance the tracking performance. Finally, we created a new roadside tracking dataset for experimental studies using Velodyne VLP-16 LiDAR sensors. The results showed that our algorithm achieved the highest accuracy and the strongest anti-occlusion performance compared with other popular algorithms such as simple online and realtime tracking (SORT), observation-centric SORT (OC-SORT), and ByteTrack. It is therefore a very suitable method for obtaining complete vehicle tracks using roadside LiDAR technology. The dataset and source code are available at https://github.com/moxigua/LMAT.

Automated summary

This article presents a lane-level and full-cycle multivehicle tracking (MVT) method using roadside LiDAR technology. A lane-level map is created by analyzing multiple frames of traffic object detection results, and an association method based on a search process and a microscopic motion model is introduced. A detection optimization approach using the lane-level map is proposed to enhance tracking performance. Experimental results demonstrate the algorithm's superior performance in anti-occlusion and accuracy.