3d Gesture recognition through RF sensing

Alejandro Alanis, Trang Thai, Gerald Dejean, Ran Gilad-Bachrach, and Dimitrios Lymberopoulos


Human interaction with devices is constrained to the surface of these devices through widely used touch sensors. In this work, we enable touchless interfaces that allow humans to interact with devices from a distance. Our approach is based on the design of a two-dimensional array of RF sensors specifically designed to detect the proximity of human body. Each sensor in the array acts as a near-field RF proximity sensor. When parts of the human body come to close proximity to the sensor, they slightly disturb its frequency response, allowing the detection of human fingers or hands. Since our approach is RF-based, it presents several distinct advantages over current sensing technologies which include the ability to work without line of sight, the ability to be easily embedded behind any type of surface, and the ability to scale to almost any size; all while operating at a similar power domain to current proximity sensing technologies. Using a prototype implementation and data collected through a user study, we demonstrate that the RF-array can detect the position and distance of a human hand located at a distance of up to 2 inches with higher than 75% accuracy.


Publication typeTechReport
PublisherMicrosoft Research Technical Report
> Publications > 3d Gesture recognition through RF sensing