Wearable Haptics: Edge Sensing and Classification of Movement for Vibratory Feedback

Abstract

Vibrotactile feedback has a variety of applications related to human movement. It has been seen to enhance motor skill learning by adding real-time tactile cues to the usual verbal and visual instruction methods used, reinforcing learning without distracting the student. In the medical field, haptic feedback is used both to augment patients’ senses (for example, as sensory substitution in prostheses) and as a targeted therapy. In people with spasticity (PwS, 12 million globally), vibration can be used to stimulate the periphery of the nervous system and relieve symptoms that can affect motor control. There are few devices that have been developed with this specific purpose in mind, and even fewer that can provide vibratory stimulation in a portable, wearer-friendly, robust manner. This ES100 capstone project focused on designing and building a wearable that could sense and classify movement to deliver relevant vibrotactile feedback for spasticity modulation. This was done through two prototypes: the first developed a signal processing protocol, and the second created an edge machine learning model to detect upper limb movements and built a wearable to deploy the model and use it to deliver targeted vibratory stimulation. The final device shows the potential of incorporating vibrotactile feedback into wearables for therapeutic purposes.