Description du projet
La robotique pour redéfinir l’efficacité de la coloscopie
L’efficacité de la coloscopie à détecter un cancer dépend de la compétence de l’opérateur, ce qui pose des problèmes alors que la demande de dépistage augmente, en particulier dans les pays européens dotés de programmes nationaux. Dans ce contexte, le projet IRE, financé par l’UE, entend transformer la technologie traditionnelle des endoscopes. Plus précisément, il fusionnera les données relatives à l’opérateur humain, la modélisation biomécanique innovante, le retour d’information sensoriel et l’entraînement simulé de robotique molle. Le projet exploite un vaste ensemble de données de plus de 2 000 coloscopies, combinant l’expérience du monde réel avec une formation simulée sur des modèles biomécaniques. Il en résultera des robots intelligents capables de naviguer de manière autonome dans les méandres de l’anatomie humaine, élevant ainsi les standards de détection précoce du cancer.
Objectif
In Intelligent Robotic Endoscopes (IRE) for Improved Healthcare Services we envision creating intelligent robotics solutions, extending current endoscope technology with robotics control that is based on learning from currently collected human operator data, coupled with novel bio-mechanical modeling techniques, and sensory feedback as well as soft robotics phantom for training.
The challenge with colonoscopy is that the success rate of detecting cancer depends on the skills of the clinician that operates the endoscope. From a health and societal perspective, the number of colonoscopies is bound to increase as they are the only way to screen patients for early cancer detection. Many European countries have national screening programs. This is a very big market in need of improved technology.
IRE enables a new generation of intelligent robots that through data, simulation and learning can interact with the interior of a living human while communicating with a human operator. The huge variation of human anatomy and the dynamic effect of human physiology make it a complicated navigational task to use endoscopes. Entanglement, haemorrhage, and perforation risks create a critical and difficult environment to navigate autonomously in where even trained human operators meet challenges. We exploit one of the largest datasets on real-life colonoscopies with more than 2,000 operations to learn safe navigation, combined with simulated training on a population of biomechanical models of the abdominal region.
IRE boosts the design and configuration of the robotic endoscope using digital twins and simulation, and careful inclusion of clinicians will speed up the process of integration. IRE will raise the level of autonomy by building upon simulation, imaging, and learning to yield an increased interpretation and understanding of the complex real- world environments, capable of anticipating the effect of human motions, adapting and replanning to avoid entanglement.
Champ scientifique
- medical and health sciencesbasic medicineanatomy and morphology
- engineering and technologyelectrical engineering, electronic engineering, information engineeringelectronic engineeringroboticssoft robotics
- medical and health sciencesclinical medicineoncology
- medical and health sciencesbasic medicinephysiology
- medical and health scienceshealth scienceshealth care services
Mots‑clés
Programme(s)
Régime de financement
HORIZON-RIA - HORIZON Research and Innovation ActionsCoordinateur
1165 Kobenhavn
Danemark