Examining Social and Environmental Factors for Wearable Data Integrity: a Case Study of Health-CPS

Abstract Wearable devices, with their ability to sense, collect, and transmit large amounts of physiological data, have experienced a recent surge in popularity, especially in the field of health monitoring. Ensuring data integrity is critical to effectively utilizing these devices, especially in the framework of Cyber-Physical Systems (CPS), where the seamless integration of computational algorithms with physical components plays a crucial role. Building on previous insights into the different contextual scenarios and their impact on data reliability, this paper examines the impact of social and environmental factors on the integrity of data collected by wearable devices, with a focus on human-related errors, such as improper wearing, and environmental conditions such as elevated humidity. Through a human-subject experiment, we simulated both prospective influential factors in a controlled lab setting and collected acceleration and heart rate data by utilizing the Apple Watch. To examine data integrity compromises, we proposed a multi-channel anomaly detection framework, and we demonstrated high accuracy rates in detecting anomalies, achieving 88.89% for improper wearing and a perfect 100% for elevated humidity. These findings confirm that human-related and environmental conditions affect data integrity. Our research enhances the reliability and robustness of CPS by quantifying the influence of social and environmental factors on wearable data integrity and providing insights into the system design beyond the technical elements of CPS.