Demo abstract: Inviz: Low-power personalized gesture recognition using wearable textile capacitive sensor arrays

摘要

This demonstration presents Inviz, a low-cost gesture recognition system that uses flexible textile-based capacitive sensors. Gestures are recognized using proximity-based movement detection using flexible capacitive sensor arrays that can be built into the environment or placed on to the body or be integrated into clothing. Inviz provides an innovative interface to home automation systems to simplify environmental control for individuals with limited-mobility resulting from paralysis, paresis, and degenerative diseases. Proximity-based sensing obviates the need for physical contact which can result in skin abrasion which is particularly deleterious to people with limited-to-no sensitivity in their extremities. A custom-designed wireless module maintains a small form factor facilitating placement based on an individual's needs. Our system leverages a hierarchical sensing technique which facilitates learning gestures based on the individual and placement of the sensors. Classification uses just-in-time embedded computational resources to provide accurate responses while maintaining a low average power consumption, in turn reducing the impact of batteries on the form factor. To illustrate the use of Inviz in a smart home environment, we demonstrate an end-to-end home automation system that controls small appliances. We will interface our system with a home automation gateway to demonstrate a subset of potential applications. This interactive demonstration highlights the intuitiveness and extensibility of the Inviz prototype.