Indoor Multihuman Device-Free Tracking System Using Multiradar Cooperative Sensing
Wei Li, Yuan Wu, Ruizhi Chen, Haitao Zhou, Yue Yu
Abstract
Robust human tracking in indoor environments is an important feature of location-based service applications, including security surveillance, elderly monitoring, and so on. Yet, the advanced technology offered by camera-based device-free tracking systems can raise privacy concerns. To address this problem, we propose a low-cost, low-power, device-free human tracking system based on millimeter-wave (mmWave) radar that can provide rich ranging and radial velocity information. In order to achieve continuous tracking, we propose a multiradar cooperative sensing scheme; with the help of the double segmentation method, we overcome the user proximity problem that prevents multiple humans from being recognized using point cloud location information. Finally, we propose a tracking and trajectory optimization algorithm that considers both spatial information and probability distribution of moving direction to output human trajectory. Experimental results show that the proposed human tracking system provides a single-human tracking error of 8.5 cm and a multihuman tracking error of nearly 10 cm.