Les travaux de recherche menés dans le domaine des environnements virtuels pour l’apprentissage humain permettent aujourd’hui d’envisager leur utilisation à des fins d’entraînement à la prise de décision en situation de crise. Le problème de ce type d’application est qu’il nécessite pour les concepteurs et les programmeurs un travail d’écriture lourd et une programmation exhaustive de l’état du monde. Notre hypothèse est de mobiliser les principes existants du récit interactif afin de proposer un design de scénarisation qui prenne en compte les aspects temporels, critiques et complexes des actions de l’apprenant. Nous proposons une approche réflexive de l’apprentissage en fondant notre scénarisation sur l’uchronie et en intégrant des notions philosophiques comme le simulacre et le rhizome dans le parcours d’apprentissage. L’apprenant pourrait revenir en arrière au sein du même scénario initié par le formateur, et le système informatique informerait ce scénario en fonction des décisions et erreurs faites par l’apprenant. Une partie de notre contribution est la conception d’une scénarisation pour un entraînement en environnement virtuel. Il respecte la réalité des experts du cas d’application concret sur lequel s’adossent nos travaux de recherches, un projet de formation de leaders médicaux au sauvetage de blessés suite à un afflux massif de blessés. / This thesis presents a story-based approach in a simulation-based learning environment in the context of crisis management. The domain of application is the management of mass casualties in the context of medical emergencies. To efficiently train medical staff, live simulations are organized, in which an entire rescue team is re-enacted. Each simulation session takes place in a pre-established pattern and includes three distinct phases : the briefing provides a framework, the scenario puts the learner in a practice of situation, and the debriefing involves the learner reflexivity with the help of the trainer’s feedback.Those simulation sessions mobilize a lot of professionals and equipment. Therefore, they cannot be used on a large scale and research is conducted to develop those trainings in virtual environments, in which the phase of debriefing is often missing while it is a crucial part in learning. Our case of application aims at creating a virtual environment for training rescue team leaders during crisis situations. In this environment, the user is a medical team leader who interacts with a team of virtual autonomous first-aid workers. In this thesis, we propose a user experience that includes both the scenario and the debriefing part of a simulation. We focus on decision making in crisis management. We suggest that the learner should reconsider his decisions if he becomes aware of a mistake he did not avoid: he could change what he has done. In the debriefing, he could visualize the consequences of those errors : he could see what he had not seen. Reflexivity of learning is an integral part of our approach. We propose a reflexive approach to learning by grounding our scenarisation on uchronia, a fiction genre consisting in stories containing "what if" scenarios at crucial points in history and present an outcome of events alternative to historical record. Our game design is based on the uchronia : when the learner becomes aware of his mistakes, he has the opportunity to go back and make a new decision. Our system will change the plot of the story by changing the occurrence probabilities of events based on a past choices of the learner, but without changing the frame or the characters in the story. Our game design offers two phases of simulation. During the scenario, the learner will explore several possible worlds. During the debrieng, he may view several possible alternatives to the scenario as a timeline.
| Identifer | oai:union.ndltd.org:theses.fr/2018COMP2415 |
| Date | 04 April 2018 |
| Creators | Laurent, Mélody |
| Contributors | Compiègne, Lourdeaux, Domitile, Szilas, Nicolas |
| Source Sets | Dépôt national des thèses électroniques françaises |
| Language | French |
| Detected Language | English |
| Type | Electronic Thesis or Dissertation, Text |
Page generated in 0.0024 seconds