BodyCompass: Monitoring Sleep Posture with Wireless Signals
summary
A wall-mounted low-power RF device emits signals, isolates RF reflections filtered by breathing to separate body echoes from static multipath, and feeds them to a deep neural network that outputs a continuous body angle (0 supine, ±90 sides, 180 prone). A deployed prototype collects 224 nights from 26 homes and achieves 87% average accuracy (94% with one week, 87% with one night, 84% with 16 minutes of labeled data) and tracks posture shifts with ~10% relative error.
AI-written summary · about
related talks
01h 13mHealth Monitoring with Wireless Sensors and Machine Learning
05mRF Vital Sign Sensing Under Free Body Movement
19mInTheMoodForLife: Analyzing Sleep to Better Manage Mood Disorders
06mWR-Hand: Wearable Armband can Track User's Hand
13mUltraPoser: Pushing the Limits of IMU-based Full-Body Pose Estimation with Ultrasound Sensing on ...
06mUnlocking the Beamforming Potential of LoRa for Long-range Multi-target Respiration Sensing
11mBaroPoser: Real-time Human Motion Tracking from IMUs and Barometers in Everyday Devices
22mExtracting Gait Velocity and Stride Length from Surrounding Radio Signals