Driver Behavior Analysis Methods: Applications oriented study

Citation:

Zinebi K, Souissi N, Tikito K. Driver Behavior Analysis Methods: Applications oriented study, in The 3rd International Conference on Big Data, Cloud and Applications - BDCA'18. Morocco ; 2018.

Abstract:

The main objective of this paper is to provide a study on driver behavior analysis methods. We focus on driver-oriented applications, with its three main sub-applications: Accident prevention, Driving styles assessment, and Driver intent prediction. The methods we are reviewing in this paper are classified according to their objective which is one of this sub- applications, and their input factors taken into account during the analysis phase. They can be either quantitative or qualitative factors. The results show that Descriptive statistics and Bayesian classifiers are the methods that were adopted in all of three sub- applications and operated on both quantitative and qualitative factors. As for the most employed methods we find Hidden Markov Model, Support Vector Machine and Image processing.