Adaptive Mobile Sensing: Leveraging Machine Learning for Efficient Human Behavior Modeling; Examining Data Privacy Regulation to Protect Personal Health Records
Perry, Lauren, School of Engineering and Applied Science, University of Virginia
Barnes, Laura, EN-Eng Sys and Environment, University of Virginia
Baritaud, Catherine, EN-Engineering and Society, University of Virginia
The technical team investigated multiple data sampling methods to determine which one most efficiently collected mobile sensing data capable of identifying user context. To do this, the team used a mobile crowdsensing application to collect data from smartphone sensors in three sensing strategies that differed in their sampling frequencies. The strategies also administered surveys to serve as ground truth data for context information such as user activity, user location, physical state, and phone position. The team created context models from the data collected in each strategy to understand the utility of the data.
The Capstone team found that the dynamic adaptive sensing strategy was the most efficient method of data collection. The dynamic adaptive sensing strategy had the best balance of reducing battery consumption while maintaining data quality by smartly turning the smartphone sensors on and off based on the phone’s use. Additionally, the data collected in the dynamic adaptive sensing strategy generated better performing context models than the data collected in the other sensing strategies. The dynamic adaptive sensing strategy will be implemented for future development of the mobile health application.
Electronic Personal Health Record systems provide patients the ability to manage their own health on personal computers or smartphones. However, some private companies creating these platforms are not under the Health Insurance Portability and Accountability Act, leaving sensitive information unprotected from being shared or misused. The gap in regulation led to researching how the United States can protect sensitive health information in these platforms. The STS research identified gaps in regulation using government articles outlining the current health data privacy regulation, news articles exposing misuse of health information in these platforms, and employing Actor Network Theory to determine the relevant actors and actants involved in the development of the platforms. Additionally, a global review of data privacy regulations served to compare current regulations in countries around the world.
Implementation of technology in the health industry has the potential to significantly improve patient health outcomes. However, there are still developments that need to be made in the technical and privacy aspects for this technology to be widely accepted. Once these advancements are made, patients will be able to safely and efficiently monitor their health from their personal devices.
BS (Bachelor of Science)
Adaptive Sensing, Actor Network Theory (ANT), Data Privacy Regulation
School of Engineering and Applied Science
Bachelor of Science in Systems Engineering
Technical Advisor: Laura Barnes
STS Advisor: Catherine Baritaud
Technical Team Members: Erin Barrett, Cameron Fard, Hannah Katinas, Charles Moens, Blake Ruddy, Shalin Shah, Ian Tucker, Tucker Wilson
All rights reserved (no additional license for public reuse)