Recent years has seen tremendous progress in the capabilities of WiFi sensing systems. Using commercial off-the-shelf devices, such as routers or phones, WiFI signals can be used to make accurate predictions about their environments. The holy grail among all sensing methods is the 3d reconstruction [1] of people that move between WiFi devices. Powerful new machine learning techniques have renewed hope in making this endeavor possible. However, such approaches need data; lots of it, precisely labeled.

At WISE and SEEMOO, we've developed a system for capturing multi-modal data—WiFi sensing, video, and sub-millimeter motion capture. If you're intrigued by the prospect of shaping high-quality datasets, delving into data processing intricacies, and contributing to meaningful research, this thesis opportunity offers a chance to be part of the evolving landscape of wireless sensing. Join us in exploring the potential of multi-modal sensing and redefining the boundaries of what's achievable!


[1]: Geng, Jiaqi, Dong Huang, and Fernando De la Torre. "DensePose From WiFi." arXiv preprint arXiv:2301.00250 (2022).