SensorHub: Multimodal Sensing in Real-Life Enables Home-Based Studies

Area of Science:

• Biomedical Engineering
• Wearable Technology
• Observational Studies

Background:

• Observational studies require real-world data collection, often involving multiple sensors.
• Integrating data from different manufacturers' sensors presents significant technical challenges.
• Existing systems struggle with simultaneous data acquisition from diverse wearable devices.

Purpose of the Study:

• To introduce SensorHub, a novel system for multimodal sensor data collection from various wearable devices.
• To enable simultaneous data acquisition from different manufacturers' sensors under real-world conditions.
• To facilitate the integration of ecological momentary assessments (EMAs) for user feedback.

Main Methods:

• Developed SensorHub to collect data from diverse wearable sensors (IMUs, ECG, EEG, PPG, EDA).
• Integrated ecological momentary assessments (EMAs) for collecting user feedback via questionnaires.
• Conducted trials with up to 21 participants, recording data at sampling frequencies up to 1000 Hz.

Main Results:

• Successfully recorded multiple signals from different devices simultaneously in a pilot study.
• Demonstrated the system's capability to handle high sampling frequencies without significant data loss.
• Validated SensorHub's effectiveness in enabling interoperability of consumer-grade sensing hardware.

Conclusions:

• SensorHub provides a robust framework for multimodal sensor data collection in real-world settings.
• The system overcomes technical difficulties in integrating diverse wearable devices for observational research.
• Enables comprehensive data capture and user feedback, advancing the feasibility of at-home studies.