LiDAR-Track: Multi-Person Positioning and Tracking Using LiDAR

Indoor multi-person positioning and trajectory tracking are crucial for applications such as smart homes, smart healthcare, and industrial security. Existing solutions suffer from low accuracy and privacy concerns, particularly in complex scenarios involving occlusions and dynamic interference. In this paper, we develop LiDAR-Track, a positioning and tracking system leveraging 2D LiDAR technology to address these issues. We propose a robust human detection and positioning method to handle occlusions and incomplete raw LiDAR readings. We also develop an efficient trajectory segmentation algorithm that leverages temporal and spatial characteristics for reliable multi-person tracking. Extensive real-world experiments demonstrate the superior performance of LiDAR-Track, achieving average positioning and single-user tracking errors of 3.01 cm and 3.05 cm, respectively. Even when tracking seven users in a complex indoor environment, the error remains within 6.55 cm.