DESIGN OF WEARABLE APPLICATIONS AND PLATFORMS FOR NEXT GENERATION PERSONALIZED SENSING

Abstract

Wearables have become increasingly ubiquitous with around 200 million smart-watch users (as of 2023). Due to battery and form-factor constraints, future wearable sensing technologies must be low-power. This prompts the need for highly interactive, low-power, and real-time next-generation wearable technologies that provide wearables with novel sensing capabilities. This thesis proposes advancements for wearable devices in two broad aspects - (a) Novel health monitoring applications that leverage existing sensors in today’s ubiquitous wearable platforms - Smartwatches and Earables, and (b) Efficient Platform optimizations for emerging wearable devices - smart textiles as well as smart AR glasses. Firstly, we showed how PPG sensors equipped in commercially available smartwatches and earphones can be leveraged for continuous molecular monitoring (pH Watch) through raw biofluids such as sweat, tears, etc. and practical, user-friendly, non-invasive in-ear blood pressure monitoring (StereoBP). Secondly, we demonstrated the first-ever on-body AI computing platform for future smart textiles comprising several on-skin interfaces (AI-on-skin) and proposed a cost-effective, low-power, and easily integrable hardware platform (RetroSphere) for low-power 3D tracking with future AR glasses.