Adaptive Passive Human Tracking with Multiple WiFi Links

Multi-link technology improves the accuracy and robustness of target tracking by leveraging the collaborative work among multiple sites or sensors. To optimize multi-link tracking performance, we design and implement an adaptive passive tracking system. Firstly, we iteratively estimate four-dimensional parameters of all dynamic paths. Secondly, we devise a dynamic link selection scheme based on parameters features. Finally, we realize adaptive passive human tracking based on parameters information of the selected links. Extensive experimental results show that our system achieves a tracking error at ${6 7 \%}$ being 0.51 m using the dynamic link selection method.