Habilitation thesis of Jérémie Dequidt

Data-driven interactive simulation for medical and robotic applications

The research works briefly reviewed in this dissertation deal with questions related to the use of data in interactive simulations for applications in medecine and soft robotics. Geometrical modeling, (bio-)mechanical modeling as well as optimization techniques to reduce the computational footprint are presented. Using patient-specific or environment data with optimization techniques and simulation enables more realistic and predictive models and opens the way for new applications. Several examples are presented in the manuscript where our simulations were fed with data from medical images, videos streams or pre-operative data in order to provide more accurate medical simulations to provide better diagnoses or in the context of soft-robotics where visual servoing provides a more robust and stable control of soft-robots.

defended on 10/02/2023