Abstract: Wearable devices of all kinds are becoming increasingly popular. One problem that plagues wearable devices, however, is how to interact with them. In this paper we construct a prototype electromyography (EMG) sensing device that captures a single channel of EMG sensor data corresponding to user gestures. We also implement a machine learning pipeline to recognize gestural input received via our prototype sensing device. Our goal is to assess the feasibility of using a BITalino EMG sensor to recognize gestural input on a mobile health (mHealth) wearable device known as Amulet. We conduct three experiments in which we use the EMG sensor to collect gestural input data from (1) the wrist, (2) the forearm, and (3) the bicep. Our results show that a single channel EMG sensor located near the wrist may be a viable approach to reliably recognizing simple gestures without mistaking them for common daily activities such as drinking from a cup, walking, or talking while moving your arms.
PDF (5551K)The copy made available here is the authors' version; for a definitive copy see the publisher's version described above.