TR2020-158

Human Pose and Seat Occupancy Classification with Commercial MMWave WiFi


    •  Yu, J., Wang, P., Koike-Akino, T., Wang, Y., Orlik, P.V., "Human Pose and Seat Occupancy Classification with Commercial MMWave WiFi", IEEE Global Communications Conference (GLOBECOM), December 2020.
      BibTeX TR2020-158 PDF
      • @inproceedings{Yu2020dec,
      • author = {Yu, Jianyuan and Wang, Pu and Koike-Akino, Toshiaki and Wang, Ye and Orlik, Philip V.},
      • title = {Human Pose and Seat Occupancy Classification with Commercial MMWave WiFi},
      • booktitle = {IEEE Global Communications Conference (GLOBECOM)},
      • year = 2020,
      • month = dec,
      • url = {https://www.merl.com/publications/TR2020-158}
      • }
  • MERL Contacts:
  • Research Areas:

    Communications, Computational Sensing, Machine Learning, Signal Processing

Our previous studies introduced a mid-grained intermediate-level channel measurement — spatial beam signalto-noise ratios (SNRs) that are inherently available and defined in the 60-GHz IEEE 802.11ad/ay standards — for the fingerprinting-based indoor localization. In this paper, we take one step further to use the mid-grained channel measurement for human monitoring applications including human pose and seat occupancy classifications. The effectiveness of the mid-grained channel measurement is validated by an in-house experimental dataset that includes 5 separate data collection sessions using classical classification methods and modern deep neural networks. Our preliminary result shows that mmWave beam SNRs are capable of delivering high classification accuracy above 90%.

 

  • Related News & Events