MAT - Monitoring Attention Trainer

Driving training using simulators has become an essential learning tool. It allows to approach pedagogical situations that are difficult to reproduce with a vehicle in an urban or extra-urban environment. During a simulator training, the quality of the interaction between the student and the instructor is fundamental, but limited by the instructor's ability to interpret the driver's attitude in real time and in post-analysis.
MAT allows the analysis and interpretation of trainee attitudes in a simulated environment.

This project is based on patented research at the crossroads of artificial intelligence and cognitive psychology carried out at INRIA, more precisely on a system of sensors and algorithms allowing to model the visual memory of a learner:

• the driver's attention (what he looks at, the attention he pays to objects, his gaze movements)

• the driver's ability to react to the scripted situations

• the level of expertise of a learner.

Analyze in a fine way the driving simulation by a student and be able to modify his training course accordingly.
Evaluate the training with cognitive scores of performance and perception of objects according to their importance in the scenario.

The prototype was tested on a professional driving simulator.

The simulators targeted are those for driving all types of vehicles (road, rail, for construction equipment ...).