WiTrack+: Modeling and Compensating Through-Wall Doppler Frequency Shift Offsets for Non-Line-of-Sight Wi-Fi Tracking

Device-free human tracking using commodity Wi-Fi has garnered significant interest due to its potential for ubiquitous indoor sensing without the need for user instrumentation. However, achieving precise tracking in non-line-of-sight (NLoS) scenarios, particularly in through-wall environments, remains a challenging issue. This is primarily due to the significant impact of wall refraction on Doppler Frequency Shift (DFS)-based tracking models, leading to accumulation of DFS offsets and subsequent trajectory drift. In this paper, we conduct a systematic theoretical analysis of the impact of wall refraction on tracking performance, aiming to uncover the quantitative relationships between DFS offsets and target motion parameters. To compensate the DFS offsets caused by walls, we propose a compensation mapping method (CompMap) that adaptively corrects DFS offsets based on system deployment and environmental parameters, eliminating the need for real-time wall refraction estimation. Furthermore, we develop WiTrack+, a robust Wi-Fi tracking system for through-wall scenarios. Extensive experiments with commodity hardware demonstrate that WiTrack+ substantially improves the tracking accuracy and stability, which validates the efficacy of the proposed compensation mechanism in suppressing the cumulative impact of DFS offset, thereby reducing path drifts and mitigating error propagation.