Personalized mHealth: Control Systems for mHealth Interventions

Designed and implemented a data-driven framework to personalize physical activity goals using system identification and three-degrees-of-freedom Kalman-filter based hybrid model predictive control (3DoF-KF HMPC) for a mobile health (mHealth) application with human in the loop. Built participant-specific behavioral models from smartwatch data and delivered adaptive step goals and rewards via a mobile app. Simulated closed-loop interventions under uncertainty using Monte Carlo methods, demonstrating robust performance and dynamic goal adjustment. This work was essential in the development and implementation of the first of its kind NIH-funded closed-loop preventative medicine intervention clinical trial, under the name of YourMove.