Wireless CSI-based head tracking in the driver seat

Xiufeng Xie, Kang G. Shin, Hamed Yousefi, Suining He

Abstract: In recent years, augmented reality (AR) has drawn significant attention in the automotive industry and shown great potential for a variety of driver-assistance applications. Tracking the driver's head is vital to seamlessly merge the AR content in the driver's view but still remains an open problem. Specifically, most existing in-vehicle AR solutions rely on cameras for head tracking, which suffer from low sampling rate, weak performance at night, and even high computation costs. Wearing a dedicated headset, on the other hand, is intrusive and inconvenient for daily driving. To overcome these limitations, we propose ViHOT, a novel wireless CSI-based predictive & device-free head tracking system for in-vehicle use. Given that drivers usually mount phones on the dashboard for navigation, ViHOT leverages the CSI of the phone's WiFi signal to track the driver's head, with a light-weight design suited for real-time driving assistance. Thanks to the high WiFi frame rate, ViHOT achieves more than 10x sampling rate over conventional camera-based approaches and thus eliminates motion blur. Moreover, ViHOT's novel tracking algorithm accurately translates CSI phase readings to head orientations (merely 4° -10° median error) without relying on any head-mounted device.