Natural Interactions, Knowledge, Immersive system for Training in Aeronautic – NIKITA

A real need of technological tools for training in professional aspects appears. It is the case in particular in the field of the aeronautics where the simulations without risks for the product and the learner, and the adaptability and the personalisation of tools would be main advantages. The virtual reality, the artificial intelligence and the serious game allow today to propose successful tools to adress this need. But two criteria ask for a reflection and for a more pushed validation: the certification of the trainings and their cost. We are going to be interested for it in three essential questions: 1) which interactions to propose? 2) which educational contents, which knowledge to represent and which reasonings to propose on these knowledge? 3) which coupling between didactic, educational and playful approaches of serious games?
Our solutions will be based on approaches in cognitive ergonomics, in knowledge engineering, in artificial and really virtual intelligence.
From the three previous questions, we have identified a major scientific lock: which contents and which interactions to improve the learning? From this scientific lock, two technical locks appear 1) the absence of models and formalisms to build and adapt dynamically the environments, the activity support in a coherent way with the aimed objectives. Or still how to conceive and to dynamically scripting at once credible virtual environments compared to the expected behavior, endowed with didactic and integral contents of the rewarding aspects to facilitate the commitment of the learner in its learnings? And 2) how make credible the gesture?
Concerning the first lock, the approach proposed by Heudiasyc UMR6599 and the LATI will be based on techniques of representation of the knowledge to allow the experts to supply models and in the decision-making modules to argue and to draw all the links of causes with effects. The originality of our approach to conceive these virtual environments for the human learning is to propose an architecture based on knowledge which allows to organize, to manage and to represent the knowledge necessary for the learning. This approach is current in the EIAH, but it is not still in the virtual reality domain. And this domain asks new questions on the more free interactions, in particular in the case of natural interactions.
Concerning the second lock, The CEA LIST has already developed an original solution which allows users to pilot their physicalised virtual in a physical simulation by using a system of motion capture. Within the framework of this project we wish to spread the proposed solution of the case of flight cameras. The originality of our approach to determine the posture of the operator is to combine a sensor which supplies directly real time a cloud of calibrated points and a physical model of the person to capture real time. Moreover the physical model which represents the learner is remote-controlled in order to take into account the biomechanical aspects and the motor control. This approach should bring a big robustness to the eclipses or in the conditions of lighting.
The industrial and economic effects can be envisaged under two aspects:
- The offer NIKITA will meet an important industrial market because the training in the maintenance upstream to the construction of devices in the aeronautical domain is one of big problems of the industry. It concerns very directly AIRBUS training, Eurocopter and of numerous sites of training - the possibility of certification in virtual environment will be a quality factor of the educational contents and the interaction modes.