Enabling Versatile Human Sensing through Wi-Fi Beamforming Feedback
Wi-Fi sensing enables contactless human sensing, while most existing systems rely on Channel State Information, which is accessible only on a limited range of Wi-Fi chipsets. In contrast, Beamforming Feedback Information (BFI) is standardized in modern Wi-Fi and readily accessible on a wide range of commodity devices, offering new opportunities for human sensing. In this paper, we present BeamLens, a Wi-Fi sensing system that extracts fine-grained spatial information from BFI to enable versatile human sensing. Specifically, we jointly estimate the angle of departure and relative time of flight from BFI to extract fine-grained spatial information and leverage two AP-STA links to capture complementary azimuth and elevation views. We further design a sparse physical antenna array and construct a uniform virtual antenna array to enhance spatial resolution. Moreover, we fuse complementary views through self-supervised learning, combining temporally aware cross-view contrastive learning and temporal masked reconstruction to learn a versatile backbone for diverse human sensing tasks. Extensive evaluations demonstrate that BeamLens achieves 90.16\% accuracy in human activity recognition, a mean per-joint position error of 8.64\,cm in 3D human pose estimation, and 70.64\% intersection over union in human silhouette estimation, while exhibiting strong generalization to unseen environments and subjects.
Authors
- Peiliang Wang (ORCID: https://orcid.org/0009-0002-6509-5875)
- Yitong Shen (ORCID: https://orcid.org/0009-0007-7420-9594)
- Mingzhe Chen (ORCID: https://orcid.org/0000-0002-2074-9487)
- Jingzhe Zhang (ORCID: https://orcid.org/0009-0007-7762-2800)
- Ning Wang
- Yili Ren
Publication Details
- Journal
- arXiv (Cornell University)
- Published
- 2026-11-12
- DOI
- https://doi.org/10.48550/arxiv.2611.16185
- Primary Topic
- Indoor and Outdoor Localization Technologies
- Type
- preprint