FingerDraw: Sub-wavelength Level Finger Motion Tracking with WiFi Signals
summary
FingerDraw reconstructs air-drawn trajectories using commodity WiFi by forming a CSI-quotient between two antennas to cancel amplitude and phase noise, then maps quotient arcs to reflection-path-length changes to extract sub-wavelength displacements. The implemented system uses one WiFi transmitter and two receivers (two-antenna RX), segments motion into circular arcs, recovers 2D trajectories iteratively, and reports a median tracking error of 1.27 cm and 93% recognition for ten digits.
AI-written summary · about