Model Predictive Control Strategies for Optimized mHealth Interventions for Physical Activity.

Area of Science:

• Behavioral Science
• Control Theory
• Health Informatics

Background:

• Many individuals do not meet recommended physical activity (PA) levels, impacting public health.
• Personalized interventions are crucial for improving adherence to PA guidelines.
• Existing behavioral strategies often lack optimization and personalization.

Purpose of the Study:

• To investigate Model Predictive Control (MPC) for delivering optimized, personalized physical activity (PA) interventions.
• To evaluate the effectiveness of MPC strategies in simulated scenarios based on the Social Cognitive Theory.
• To explore the interplay of measurement, information, and decision-making in PA interventions.

Main Methods:

• Utilized a health behavior fluid analogy model grounded in Social Cognitive Theory.
• Evaluated diverse intervention strategies within simulated scenarios.
• Employed constrained Model Predictive Control (MPC) controllers for optimization.

Main Results:

• MPC demonstrates potential for feasible, personalized, and user-friendly PA interventions.
• Effectiveness of MPC was shown even with limited measurement data.
• Constrained MPC controllers proved effective in optimizing PA interventions in simulations.

Conclusions:

• Model Predictive Control (MPC) offers a promising framework for personalized physical activity (PA) interventions.
• The study provides a foundation for real-world experimental evaluation of MPC strategies.
• Optimized behavioral interventions can enhance public health by increasing PA levels.