Nonparametric Hammerstein model based model predictive control for heart rate regulation.

Area of Science:

• Biomedical Engineering
• Control Systems
• Cardiovascular Physiology

Background:

• Human cardiovascular system modeling is complex due to unknown structures.
• Nonparametric modeling offers a realistic approach for complex biological systems.
• Accurate heart rate regulation is crucial for safe and effective treadmill exercise.

Purpose of the Study:

• To develop a novel nonparametric model predictive control (MPC) approach for heart rate regulation during treadmill exercise.
• To address the challenges of modeling the human cardiovascular system using nonparametric methods.
• To design a controller that ensures safe exercise by adhering to speed and acceleration constraints.

Main Methods:

• A new nonparametric Hammerstein model identification approach was developed for heart rate response.
• Decoupled identification of linear dynamics and input nonlinearity using pseudo-random binary sequence data.
• Correlation analysis for linear component step response and Support Vector Regression for inverse static nonlinearity.
• Model predictive controller designed with predefined speed and acceleration constraints.

Main Results:

• The proposed method successfully identified the nonparametric Hammerstein model for heart rate dynamics.
• The developed MPC algorithm achieved optimal heart rate tracking.
• Simulations demonstrated effective performance under specified exercise constraints.
• The control strategy enhances safety during treadmill workouts.

Conclusions:

• The novel nonparametric MPC approach provides an effective solution for heart rate regulation during exercise.
• The identification method accurately models complex cardiovascular dynamics.
• This approach contributes to safer and more optimized treadmill exercise protocols.