TWIWT XXXII: Week of May 19, 2025
More wearable technology news, this week featuring Android XR announcement + smart glasses powered by Gemini AI, wearables for hospital mobility monitoring, motion tracking, and more.
Google demos AndroidXR on hardware
Google has launched Android XR, its first Android platform built for the Gemini AI era, supporting devices from immersive headsets to lightweight smart glasses. Developed with Samsung and optimized for Snapdragon processors, the platform features Samsung's Project Moohan headset launching later this year and Android XR glasses equipped with cameras, microphones, speakers, and optional in-lens displays. The glasses provide hands-free AI assistance through Gemini integration, enabling text messaging, navigation, real-time translation, and visual search without requiring phone interaction. Google has partnered with Gentle Monster and Warby Parker for consumer eyewear production, while hundreds of developers are building applications following the developer preview release, positioning smart glasses as a natural interface for ambient computing.
ActiGraph wearables validated for hospital mobility monitoring in older adults
A comprehensive feasibility study at an academic hospital in Hamilton, Ontario has identified the ActiGraph wGT3X-BT as the optimal wearable device for monitoring mobility in hospitalized older adults. The research evaluated four consumer and medical-grade devices across 55 participants aged 60 and older, addressing the critical issue that hospitalized patients typically spend only 3% of their time standing or walking, with each day of bed rest causing 1-5% muscle strength loss. The ActiGraph demonstrated superior performance with 100% data retrieval compared to 72-81% for competing devices, while ankle-worn configurations achieved excellent step counting reliability (ICC=0.94) and thigh placement accurately detected sedentary behavior with 91% standing position accuracy. Despite 96% of participants having no prior wearable experience, 80% expressed willingness to wear monitoring devices during hospitalization, though practical challenges included device removal due to skin irritation and swelling. The findings support implementing dual-device configurations combining thigh and ankle placement to capture comprehensive mobility data, providing evidence for wearable technology as an early intervention tool to prevent functional decline in older hospitalized patients.
You can read the full paper at JMIR Aging.
Machine learning model enhances wearable motion tracking with behavioral constraints
Harvard researchers have developed the Activity-in-the-loop Kinematics Estimator (AIL-KE), a machine learning system that significantly improves motion tracking accuracy using just two inertial measurement units by incorporating activity classification as a behavioral constraint. The system integrates three neural network components that identify user activities and apply appropriate motion constraints, addressing the fundamental challenge of IMU sensor drift that accumulates over time. Testing across strength training exercises and simulated industrial work demonstrated substantial performance improvements, with trajectory errors under 0.021 meters (52% better than baseline systems) and shoulder joint angle errors below 6.5 degrees over 10-minute periods. Unlike existing approaches requiring extensive body measurements or numerous sensors, AIL-KE operates with minimal hardware while maintaining accuracy comparable to laboratory-grade optical motion capture systems, with potential applications in strength training analysis, ergonomic assessment, rehabilitation monitoring, and integration into commercial wearable devices.
You can read the full paper here. Credit: Dmitry Popov.
CorePower Yoga integrates Oura Ring biometrics for personalized class recommendations
CorePower Yoga, the largest yoga studio chain in the United States, has launched a health data integration with Oura's smart ring technology that delivers personalized workout recommendations based on biometric feedback. The partnership, live as of May 21, incorporates Oura's Readiness Score—which synthesizes sleep quality, heart rate variability, and stress indicators—into CorePower's mobile app to suggest appropriate class intensity levels based on members' recovery status. The system recommends high-intensity sessions for optimal recovery metrics while directing users with lower readiness scores toward restorative classes. Cross-promotional offers include 10% discounts on Oura's fourth-generation ring for CorePower members and 15% off unlimited memberships for Oura users, reflecting the growing trend of wearable-guided fitness programming that moves beyond activity tracking toward predictive wellness guidance.
Jobs and other opportunities
Other interesting links and stuff missed from previous weeks
A full upper-body biomechanical model that can learn to reach.
SageMotion now supports real-time EMG sensing + haptic/audio/visual biofeedback.
Dose-response relationship between evening exercise and sleep.
AutoPeel: the first autonomous robot approach for chronic wound dressing.