Overview
Our research develops AI-powered, wearable smart glasses that monitor daily activities and detect critical transitions to improve safety, independence, and health outcomes for older adults, stroke patients, and visually impaired users. By combining inertial sensors and machine learning, the glasses provide real-time alerts to prevent falls and support human-centered interventions.
Motivation
- Falls are a major health risk for older adults and patients with visual impairments or mobility challenges.
- Improper use of single or multi-focal glasses can impair depth perception, increasing fall risk.
- There is a need for unobtrusive, real-time monitoring solutions that preserve user independence.
Our Approach
- ActiSenSee: Glasses with inertial sensors detect discrete and continuous activities and critical transitions (e.g., sitting-to-standing) with up to 95% accuracy.
- Activisee: Sensor-augmented glasses detect activity transitions and alert users to switch eyewear when needed, achieving up to 92% accuracy.
Both systems use machine learning to convert raw sensor data into actionable real-time alerts, enabling human-centered safety interventions.
Key Results
- High-accuracy detection of discrete and continuous activities using smart glasses.
- Reliable detection of activity transitions associated with fall risk.
- Real-time user alerts prevent unsafe behavior and improve safety.
- Demonstrated potential for home and clinical deployment.
Impact
- Prevention of falls due to visual distortion or improper eyewear use.
- Support for older adults, stroke patients, and children with disabilities.
- Integration of human-centered AI with wearable technology for real-time health monitoring.
- Foundation for future multimodal sensing and predictive health interventions.