Offset-free MPC strategy for nonzero regulation of linear impulsive systems

Area of Science:

• Biomedical Engineering
• Control Systems Theory
• Systems Biology

Background:

• Drug administration in biomedical applications is often modeled as an impulsive control system.
• Plant-model mismatch, caused by physiological changes, leads to inaccurate patient treatment.
• Existing control strategies for impulsive systems do not address offset elimination due to plant-model mismatch.

Purpose of the Study:

• To propose and validate a novel offset-free model predictive control (MPC) strategy for impulsive systems.
• To address the challenge of plant-model mismatch in biomedical control applications.
• To ensure accurate and effective drug administration by eliminating treatment offsets.

Main Methods:

• Introduction of an impulsive disturbance model.
• Development of an observer with new observability criteria for state and disturbance estimation.
• Integration of a recent MPC formulation using artificial variables for regulation.
• Validation using a dynamical model of type 1 diabetes.

Main Results:

• The proposed observer design enables accurate estimation of system states and disturbances.
• The novel MPC strategy achieves zero offset tracking under steady-state conditions.
• The controller effectively steers the system state to a desired equilibrium setpoint.
• Demonstrated efficacy in a type 1 diabetes dynamical model.

Conclusions:

• The developed offset-free MPC strategy is effective for impulsive biomedical systems.
• The proposed method successfully compensates for plant-model mismatch, improving treatment accuracy.
• This approach offers a robust solution for real-time drug administration control in dynamic physiological environments.