GIPSA-Lab , Salle JM Chassery, 11 rue des Maths
12 December 2024 - 14h00
Data-based anesthesia process modelling for online monitoring and prediction (Phd Defense)
by Bob AUBOUIN-PAIRAULT from Université Grenoble Alpes, VERIMAG, GIPSA-Lab
12 December 2024 - 14h00
Data-based anesthesia process modelling for online monitoring and prediction (Phd Defense)
by Bob AUBOUIN-PAIRAULT from Université Grenoble Alpes, VERIMAG, GIPSA-Lab
Abstract: This thesis explores computational methods to enhance the safety and effectiveness of anesthesia management, a critical yet complex aspect of modern surgery. By leveraging control theory and machine learning, the work focuses on improving the use of intravenous drugs like propofol and remifentanil to regulate the depth of hypnosis and improve the alarm system.
Key contributions include a new pharmacodynamic model integrating machine-learning regressors, the design of closed-loop controllers with advanced estimation techniques coupled with model predictive control, a filtering method to increase alarm specificity, and a comprehensive framework for development of predictive alarms in the operating room. These innovations demonstrate the potential of combining control theory and machine learning to improve anesthesia practice, paving the way for safer and more efficient clinical applications.
