Tracking Synchronous Gestures with WiFi

Tracking synchronous gestures benefits various applications in team sport training, group interactive gaming, and computer-supported cooperative working. We present WiSync, a ubiquitous synchronous gesture tracking scheme on a single WiFi link. It reuses the pervasive WiFi networks, imposes no wearable sensors on users and operates in Non-Line-Of-Sight (NLOS) propagation. The idea is that with synchronous gestures from multiple users, the received signals exhibit characteristics as if a single user were performing a different gesture, while in the case of asynchronous gestures, they demonstrate chaotic patterns. To discern synchronous and asynchronous gestures without gesture recognition, we harness their differences in the periodicity during gesture repetitions. We prototype WiSync on commodity WiFi infrastructures and evaluate it in two indoor environments. Experimental results show that WiSync achieves a balanced synchronous gesture identification accuracy of 90.43% for 2 users performing 8 types of gestures. WiSync is also robust to user orientations and can scale to 4 users while retaining a satisfactory performance.