Fabric as a Sensor: Towards Unobtrusive Sensing of Human Behavior with Triboelectric Textiles.

Smart apparel with embedded sensors have the potential to revolutionize human behavior sensing by leveraging everyday clothing as the sensing substrate. However, existing textile-based sensing techniques rely on tight-fitting garments to obtain sufficient signal to noise, making it uncomfortable to wear and limiting the technology to niche applications like athletic performance monitoring. Our solution leverages functionalized fabric to measure the triboelectric charges induced by folding and compression of the textile itself, making it a more natural fit for everyday clothing. However, the large sensing surface of a functionalized textile also increases body-coupled noise and motion artifacts, and introduces new challenges in how we suppress noise to detect the weak triboelectric signal. We address these challenges using a combination of textile, electronics, and signal analysis-based innovations, and robustly sense joint motions by improving SNR and extracting highly discriminative features from the signal. Additionally, we demonstrate how the same sensor can be used to measure relative changes in skin moisture levels induced by sweating. Our design uses a simple-to-manufacture layered architecture that can be incorporated into any conventional, loosely worn textile. We show that the sensor has high performance in natural conditions by benchmarking the accuracy of sensing several kinematic metrics as well as sweat level. Additionally, we provide real-world performance evaluations across three application case studies including activity classification, perspiration measurements during exercise, and comfort level detection for HVAC systems.