Thermal sensors improve wrist-worn position tracking

Area of Science:

• Biomedical Engineering
• Human-Computer Interaction
• Behavioral Science

Background:

• Wearable sensors (IMUs, proximity) struggle with precise hand-to-head position tracking.
• Accurate monitoring is crucial for behaviors like body-focused repetitive behaviors (BFRBs).
• BFRBs are often misdiagnosed and undertreated due to monitoring limitations.

Purpose of the Study:

• To evaluate the efficacy of incorporating thermal sensors into wearable devices for improved hand position tracking.
• To assess the Tingle device's ability to accurately distinguish between different head locations for behavior monitoring.
• To explore the potential of thermal sensing wearables in clinical applications, particularly for BFRBs.

Main Methods:

• Developed the Tingle wearable device integrating thermal sensors alongside inertial measurement units (IMUs) and proximity sensors.
• Compared tracking accuracy of thermal sensors against IMU and proximity sensors alone.
• Tested the device on 39 adult participants, differentiating behaviors across six head locations.

Main Results:

• Thermal sensors significantly enhanced position tracking accuracy compared to IMUs and proximity sensors alone.
• The Tingle device achieved high accuracy (AUROC values) in distinguishing behaviors from various head locations.
• The best performance was observed for the back of the head (median AUROC 1.0), with good performance on the cheek (median AUROC 0.93).

Conclusions:

• Integrating thermal sensors offers a significant advantage for accurate hand position tracking in wearable devices.
• The Tingle device demonstrates promising potential for precise behavioral monitoring, especially for BFRBs.
• This technology could improve diagnosis and management strategies for conditions requiring accurate behavioral tracking.