Cyber–Physical Mobile Arm Gesture Recognition using Ultrasound and Motion Data

Alternative and more intuitive input methods for the control of consumer devices (e.g. tablets, smart-phones and home automation systems), vehicle entertainment systems and industrial machines and robots are on the rise. In this contribution, we propose a body worn setup, which supplements the omnipresent 3 DoF motion sensors with a set of ultrasound transceivers. This allows for an active time–of–flight based measurement of distances between independent nodes on upper limbs and torso without requiring calibration or remote reference sensors like video cameras. A body worn demonstrator is introduced, which consists of three independent but synchronized nodes, equipped with motion sensors and Capacitive Micromachined Ultrasonic Transceivers. Results of an initial test series using eight wide arm gestures show, that combining ultrasound distance measurements and motion data improves the classification accuracy.