WiStress: Contactless Stress Monitoring using Wireless Signals
summary
A desktop prototype transmits ultra‑low‑power wireless signals with an 77 GHz RF sensor and recovers tiny body-motion, breathing, and heartbeat reflections to compute interbeat intervals using a self‑similarity machine‑learning extractor. A stress classifier combines IBI and motion features and is validated with 22 subjects, 84–95% accuracy across induced and free‑living conditions. The system operates up to 12 ft range, across orientations and activities (sleeping, reading, speaking, typing) and runs fully passively.
AI-written summary · about