Toward Device-Free Micro-Gesture Tracking via Accurate Acoustic Doppler-Shift Detection
Wenyuan Liu, Weihang Shen, Binbin Li, Lin Wang · IEEE Access · 2018
As a universal interaction method, hand and finger gestures can express people’s intention directly and clearly in daily life, and has been one of the hotspots in human–computer interaction community. In this paper, we presentDMT, a device-free finger gesture tracking system that can track and recognize the finger motion accurately. To achieve this, we transform the mobile device, such as a smart phone, into an active sonar system by establishing inaudible audio links between the built-in speakers and microphone. The finger motion will have an effect (e.g., Doppler-shift) on the audio signals, which makes it possible to track the finger motion according to the received signal characteristics at the microphone side. Due to the small reflection energy and slow moving speed of the finger, the existing methods cannot detect the Doppler-shift accurately. To this end, a Fourier fitting-based method is proposed in theDMTto accurately detect the Doppler-shift. With the detected Doppler-shift, theDMTcan track the finger motion with high accuracy. TheDMTsupports all kinds of finger gestures interaction, including characters and shapes. Extensive experiments demonstrate the high accuracy and robustness of theDMTin dynamic environments.