PhD

Vision-based Trajectory Tracking Robust to Modeling Errors

PhD Project short description

Automatic tasks in medical robotics are commonly performed in the fields of orthopedic surgery or radiotherapy, but very rarely in digestive surgery. One of the difficulties is the handling of model errors in minimally invasive surgical robots, in particular the ones caused by cable transmissions. Even in the case of movements carried out in closed loop under the feedback of an endoscopic camera, the movements are often imprecise, slow and unnatural, which strongly limits the interest of automation.

In this thesis work, we propose to develop a new paradigm for the control of robotic surgical instruments under the feedback of endoscopic cameras. Rather than trying to improve behaviors by fine modeling, we propose to integrate uncertainties on the movements of the instruments into the realization of the tasks. In return, we will accept not to carry out the task exactly by authorizing margins of precision. The general objective is to be able to achieve smoother movements while obtaining precision similar to manual control.

From the application point of view, we will be interested in laser treatment tasks in robotic flexible endoscopy. Flexible endoscopes have complex and variable behavior over time and depending on their conditions of use and are therefore very good candidates for the application of the methods that we wish to develop.

Here is the link to the complete description of the PhD proposal: https://docs.google.com/document/d/1G0mA_ciUroCLSFogS6FKxDxYnIy2Hzc5R_eNCH8T6CE/edit?usp=sharing

Working Environment

The PhD thesis will be hosted in the RDH team (Robotics and Data Science for Health) of the ICube laboratory (joint lab of University of Strasbourg and French National Center for Research (CNRS)), ( https://icube.unistra.fr/en/ ) located in the downtown hospital of Strasbourg.

The PhD work will be supervised by Florent Nageotte (Associate Pr, Habilited to direct research). The PhD will be funded for 3 years by a national Grant. There will be opportunities to teach.

Application

We are looking for a high-ranked candidate who will have completed his/her Master degree by September, with background in robotics or automatic control. Experience or knowledge in computer vision and machine learning will be appreciated but are not mandatory. Advanced skills in programming (Python or C/C++) are expected.

The selection process will take place in two steps: - First selection of candidates on the fly on the basis of provided written documents (see below) and interviews with PhD supervisors - For candidates selected after the first round, interview by a university committee on June 13 or June 14.

To apply send a CV, cover letter, master program and master grades (M1 and first semester of M2) before June 1st to: Nageotte@unistra.fr

PhD starting dates: between September and November 2023

Robot-assisted, focused ultrasound device for volumetric Blood-Brain-Barrier opening

PhD Project short description

The Blood-Brain Barrier (BBB) is a natural physiological barrier that prevents pathogens and harmful molecules from entering brain tissue. BBB also blocks large molecules, such as therapeutic drugs. In a report issued in 2005, BBB was considered to be the major bottleneck in brain drug development. Focused ultrasound, in combination with the injection of microbubbles, has the potential to open the BBB in a localized, transient and reversible manner. Except for implanted devices that are highly invasive, all existing studies on BBB opening are restricted to single-point focusing. From a medical point-of-view, BBB should ideally be open in larger volumes, such as the peritumoral region in the case of brain tumors. The most promising solution to achieve this goal is the use of robotics.

The RDH team of the ICube laboratory has been developing a robot-assisted, neuronavigated BBB opening device, in collaboration with the CEA/Neurospin, a center renowned for its contributions in the field of ultrasound-mediated BBB opening. This first prototype has been shown to allow for accurate targeting of almost any specific point in the brain, taking both acoustic and robotic constraints into account. The objective of the PhD is to develop a fully operational prototype for preclinical volumetric BBB opening.

Here is the link to the complete description of the PhD proposal: https://docs.google.com/document/d/1S37WLCT-a8ZX0NuWHzevUcGRwoAj9ubCF40KVFCs3pU/edit?usp=sharing

Working Environment

The PhD thesis will be hosted in the RDH team (Robotics and Data Science for Health) of the ICube laboratory (joint lab of University of Strasbourg and French National Center for Research (CNRS)), ( https://icube.unistra.fr/en/ ) located in the downtown hospital of Strasbourg.

The PhD student will join a multi-disciplinary team made of researchers, engineers and students working in robotics, physics or ultrasounds and medicine. The PhD work will be supervised by Florent Nageotte (Associate Pr.) and Jonathan Vappou (Research Scientist).

The PhD will be funded for 3 years by the Healthtech Institute. There will be opportunities to teach.

Application

We are looking for a high-ranked candidate who will have completed his/her Master degree by September, with background in electrical engineering or biomedical engineering. Previous experience in robotics is recommended. Advanced skills in programming (Python or C/C++) are expected. The candidate should be willing to work using a real interdisciplinary approach, i.e., his/her work will be mainly centered on robotics, but he/she should have a thorough understanding of the underlying ultrasound physics and physiology.

The selection process will take place in two steps: - First selection of candidates on the fly on the basis of provided written documents (see below) and interviews with PhD supervisors - For candidates selected after the first round, interview by a Healthtech committee end of May (dates to be defined).

To apply send a CV, cover letter, master program and master grades (M1 and first semester of M2) before May 8th to: Nageotte@unistra.fr and jvappou@unistra.fr

PhD starting dates: between September and November 2023

ÉDroPoMe (Endurance Drone for Pollution Measurement) - Utilisation d'un drone pour la cartographie 3D in-situ de nuages de fumées

Le projet vise à développer un outil de cartographie 3D rapide et temps réel de l’évolution des polluants d’un panache de fumées d’incendies, grâce à un drone de grande endurance à propulsion à hydrogène non polluante (5h d’autonomie) embarquant des capteurs de pollution. Des techniques de commande avancée permettront une planification de trajectoire en temps réel afin que le drone vole à la frontière du panache de fumée, évitant ainsi la saturation des capteurs embarqués et maximisant la pertinence des données récoltées. Les données serviront à affiner les prédictions fournies par un modèle d’écoulement des fumées à l’aide d'outils d’intelligence artificielle. Le projet débouchera sur un démonstrateur en vue d’une validation expérimentale, en collaboration avec les professionnels du métier avec lesquels nous travaillons (pompiers du SIS67).

Profil recherché : Ingénieur ou Master avec une spécialisation en automatique. Une expérience dans le domaine des drones sera appréciée.

File:EDroPoMe

Sujet détaillé