Robust PID control of propofol anaesthesia: Uncertainty limits performance, not PID structure

Area of Science:

• Anesthesiology
• Control Engineering
• Pharmacology

Background:

• Continuous development of advanced control structures for anesthesia hypnosis regulation.
• Lack of comparative studies evaluating advanced structures against clinically validated PID controllers.
• Need to analyze performance limitations of filtered PID controllers versus high-order Youla parameter controllers.

Purpose of the Study:

• To analyze the performance limitations of filtered PID controllers compared to high-order controllers (Youla parameter).
• To compare the clinical benefits of advanced control structures against simpler PID-based solutions.
• To assess the impact of inter-patient variability on controller performance.

Main Methods:

• Synthesized two robust optimal filtered PID controllers considering inter-patient variability using 47 paediatric pharmacological models.
• Synthesized individualized filtered PID and Youla controllers for each model.
• Utilized convex optimization and gradient-based methods for controller synthesis.
• Evaluated worst-case performance across the patient model set.

Main Results:

• Optimized robust filtered PID controllers and individual-specific PID and Youla controllers.
• Similar frequency response characteristics observed across all designs.
• No significant performance improvement with Youla controllers compared to individually tuned PID controllers.
• Inter-patient variability had a larger impact on performance than controller structure differences.
• Robust solutions exhibited slower responses in simulations.

Conclusions:

• Individual-specific Youla parameter and filtered PID controllers showed only marginal performance differences.
• Inter-patient variability is a more significant factor limiting performance than the structural limitations of filtered PID controllers.
• Filtered PID controllers remain a viable option when considering clinical and technical constraints.