Toothbrushing Monitoring using Wrist Watch
Toothrushing is essential for people's dental health, but the failure to comply with the best toothbrushing practice can cause excessive tooth abrasion or insufficient cleaning.
Common mistakes include under-brushing, over-brushing, and incorrect brushing techniques.
Using a smartwatch, we design a system that can conveniently detect these problems. We design an algorithm that identify the toothbrushing surfaces and techniques based on the the user's hand movements. We design a novel sensing technique that attaches a small magnet on the toothbrush handle, and use the magnetometer to recognize the toothbrush motion. In a 3-week study with 12 users, our system recognized toothbrushing gestures with an average precision of 85.6%.
Safe Driving Monitoring with Wearable Magnetics
Gloally, road accidents cause 3200+ deaths daily, and surveys show 94% of all accidents are caused by human errors. Common dangerous driving behaviors include manual distractions activities, such as eating or texting, of either hand, visual distractions, when the driver takes vision off the road and focus on roadside objects instead, drowsy driving and unsafe lane changing/turning. It's challenging to monitor driver behaviors in real time because these behaviors involve the simultaneous motions of the driver's two hands and head. It is essential to monitor all these three parts to ensure accurate driving monitoring.
In this project, we propose a novel technique attaches magnets to the user's wearable accessaries, ring and eye-classess in particular, and use the onboard smartwatch magnetometer to monitor all the driving motions. We design a novel Simultaneous Tracking and Classification algorithm that firstly identifies which body part is moving, then tracks its motions. Based on the tracking results, we enable the detection of many unsafe driving behaviors. In a roadtest study with 10 users, we achieved a detection accuracy of 89.6%.