Jérôme Cieslak’s seminar, IMS. Title of his presentation: "Sensors: modeling and identification. Application to avionics and physiological data."
In the French language, the word "modèle" has different meanings. For some people, this term highlights an example to follow. For others, it refers to a type of car or to models found on the cover of a magazine. In the physical and engineering sciences, control theory uses this word to highlight a mathematical representation of a system or a physical phenomenon. These mathematical models can be obtained by the laws of physics (we speak of knowledge model) or by using identification techniques on experimental data sets (we speak of behavior model). These models can be used as "software sensors" to access variables that cannot be measured directly. They can also be used to design a control law capable of modifying the natural behavior of a system, or be of a certain complexity to reproduce the behavior of the real system with high fidelity. In this presentation, the focus will first be on the development of model-based techniques in an industrial aeronautics validation and verification process. Then, it will be shown how models can be used in innovation to, for example, avoid the need for preclinical animal testing in the context of type 1 diabetes.
Biography of the speaker: Jérôme Cieslak is an associate professor with the University of Bordeaux, France. He specializes in control engineering. His expertise areas include Fault Detection and Diagnosis (FDD), Fault-Tolerant Control (FTC) and intelligent strategies for complex aerial systems (aircraft, satellite, UAV, …). He investigates specifically the transient management due to the interaction between the FDI and FTC units. For the last 5 years, he has also been addressing control theory related to healthcare by developing a robust artificial pancreas based on biosensor measurements, in order to provide a viable solution against the high variability of Type 1 diabetes patients.