Self-learning based motion recognition using sensors embedded in a smartphone for mobile healthcare

Abstract

Human motion recognition using wearable sensors is becoming a popular topic in the field of mobile health recently. However, most previous studies haven’t solved the problem of unlabeled motion recognition very well due to the limitation of learning ability of their systems. In this paper, we propose a self-learning based motion recognition scheme for mobile healthcare, in which a patient only needs to carry an ordinary smartphone that integrates some common inertial sensors, and both labeled and unlabeled motion types can be recognized by using a self-learning data analysis scheme. Experimental results demonstrate that the proposed self-learning scheme behaves better than some existing ones, and its average accuracy reaches above 80% for motion recognition.