Driver behavior is receiving increasing attention as a result of the staggering number of road accidents. Many road safety reports regard human behavior as the most important factor in the likelihood of accidents. The detection and classification of aggressive or abnormal driver behavior is an essential requirement in the real world to avoid deadly road accidents and to protect road users. The automatic detection of the driver's behavior aids in the prevention of dangerous situations for the driver and all other participants in the driving environment, as well as the implementation of corrective measures. This paper presents a systematic literature review (SLR) of the classification of driver behavior. The study aim is to highlight and analyze the different types of driver behavior, data sources, datasets, features, and artificial intelligence techniques used to classify driver behavior and its performance. Based on the results obtained from the analysis of the selected works, we aim to identify the key contributions and challenges of studying driver behavior classification and propose potential avenues for further directions for practitioners and researchers.INDEX TERMS Driver behavior, intelligent transport system, systematic literature review, machine learning, deep learning.

No abstract

Abstract. The violation of traffic rules is, nowadays, the most important cause of accidents. Passing an intersection or a red light can be fatal for a driver and lead to serious damage. In fact, when the driver encounters a signal change from green to yellow, he or she is required to make a decision to stop or to go based on many factors. Making the wrong decision will result in a red-light violation or an abrupt stop at the intersection. Researchers typically focus on the connection between driving behavior and decision-making because of its importance in controlling aggressive drivers’ behavior. This work aims to compare the potential of machine learning techniques to classify driver behavior at intersections and follows a data preparation process to expect interesting performance results. A comparative study was therefore conducted to explore the various data source and algorithms employed to classify driver behaviors at intersections and to address the most important techniques used. Two experiments were also developed in this paper. The first experience attempts to classify driver behavior in intersections into (1) stopping and (2) going at intersections. The second experience was based on stopping observations when approaching intersections. We classified these drivers into two categories: those who stop beyond the line (1) are considered dangerous or unsafe stops, and those who stop before the line (2) are considered safe stops. As a result, XBboost archive the best performance with 92.19% of accuracy and 94.38% of precision in the first experience and RF gives the best performance in the second experience with an accuracy of 99.38%.

No abstract