In this paper we will present how the Hidden Markov Model could be used to analyse the evolution of the sensors data regarding the driving situation. Using data from the vehicle's dynamic and from the driver's actions, we show how this type of model allows us to predict with a 85% success rate in which situation the driver is. Furthermore the statistic automatically learned model could contain some information useful for experts In this paper, firstly we will describe the interest in using this type of modelling in the field of transportation. Then we will describe the methodology we used to collect and analyse data on driver behaviour. In the last part, we will present and discuss our results. (Author/publisher)
Abstract