SensorW: Open-Source Wear Operating System Platform for Continuous Global Positioning System and Multimodal Sensor Data Collection in Real-World Settings.

Digital phenotyping requires continuous, real-world collection of multimodal behavioral and physiological data. Although several smartphone-based sensing frameworks have been developed and used, open-source platforms for long-term smartwatch sensor data acquisition remain limited. To bridge this gap, SensorW was developed, an open-source Wear operating system (OS)-based framework for direct lifelog data collection from smartwatches without requiring a paired smartphone or manufacturer cloud service. SensorW includes a front-end Wear OS application for sensor data collection and a back-end server for application initiation, secure data submission, and storage. The front-end application uses battery-management strategies, such as hybrid Global Positioning System sampling and wake-lock management, to continuously collect multimodal sensor outputs, including location data, physical activity-related metrics, barometric pressure, and device wear status, and submit data to the back-end server. A preliminary technical evaluation of SensorW was conducted with seven participants who wore SensorW-enabled devices continuously for at least seven consecutive days. The technical evaluation demonstrated multimodal data collection during 17-42 days of monitoring, with operational data availability ranging from 0.73 to 1.00 across participants. These findings support the technical feasibility of SensorW as a practical and extensible platform for continuous digital phenotyping using consumer-grade smartwatches.