Adaptive Guiding for Fluvial Navigation Training in Informed Virtual Environment
Résumé
The impact of virtual environments is now recognized for training, but systems remain unable to give meaning to real-life experiences during complex situations. Training in Informed Virtual Environment (environment including knowledge-based models and providing an action/perception coupling) allows taking the user's activity into account.We propose a system which can automatically guide learners by displaying multimodal feedbacks according to their level, their state and the situation. A decision-making module, based on the Dempster-Shafer theory, determines the most appropriate aids according to the learner's actions (mistakes made, risks taken) and his/her state (stress, cognitive load, etc.). This system allows relieving the trainer from controlling the whole training simulation. First results on the decision-making module show the potential of our approach for complex situations with uncertain data.