FingerDraw

• DOI: 10.1145/3380981
• Publication Date: 2020-03-18T18:54:31Z
• Authors: Dan Wu; Ruiyang Gao; Youwei Zeng; Jinyi Liu; Leye Wang; Tao Gu; Daqing Zhang
• Keywords: Tracking (education)

This paper explores the possibility of tracking finger drawings in air leveraging WiFi signals from commodity devices. Prior solutions typically require user to hold a wireless transmitter, or need proprietary hardware. They can only recognize small set pre-defined hand gestures. introduces FingerDraw, first sub-wavelength level motion system using devices, without attaching any sensor finger. FingerDraw reconstruct drawing trajectory such as digits, alphabets, and symbols with setting one transmitter two receivers. It uses two-antenna receiver sense scale displacement each direction. The theoretical underpinning is our proposed CSI-quotient model, which channel quotient between antennas cancel out noise CSI amplitude random offsets phase, quantifies correlation value dynamics object displacement. sensitive enables us detect changes In-phase Quadrature parts state information due movement. Our experimental results show that overall median accuracy 1.27 cm, recognition ten digits achieves an average over 93.0%.