Moving attention from the road: A new methodology for the driver distraction evaluation using machine learning approaches

This work describes an approach to develop a model of driver´s distraction induced by an on-board information system basing on vehicle data. Machine learning techniques have been adopted to find the model able to better predict distraction, given a target value. Results pointed in favor of a model obtained with the ANFIS (adaptive neuro fuzzy inference system) technique. Further investigations will be carried out by porting the model on real car.