The application of machine learning techniques for driving behavior analysis: A conceptual framework and a systematic literature review. (January 2020)

• Record Type: Journal Article
• Authors: Elamrani Abou Elassad, Zouhair; Mousannif, Hajar; Al Moatassime, Hassan; Karkouch, Aimad
• Is Part Of: Engineering applications of artificial intelligence. Volume 87(2020)
• Publication Date: 2020-01
• Subjects: Machine learning -- Driving behavior analysis -- Conceptual framework -- Systematic literature review; Engineering -- Data processing -- Periodicals; Artificial intelligence -- Periodicals; Expert systems (Computer science) -- Periodicals; Ingénierie -- Informatique -- Périodiques; Intelligence artificielle -- Périodiques; Systèmes experts (Informatique) -- Périodiques; Artificial intelligence; Engineering -- Data processing; Expert systems (Computer science); Periodicals; 620.00285
• DOI: 10.1016/j.engappai.2019.103312
• Languages: English
• ISSNs: 0952-1976

Abstract

Abstract: Driving Behavior (DB) is a complex concept describing how the driver operates the vehicle in the context of the driving scene and surrounding environment. Recently, DB assessment has become an emerging topic of great importance. However, in view of to the stochastic nature of driving, measuring and modeling, DB continues to be a challenging topic today. As such, this paper argues that to move forward in understanding the individual and organizational mechanisms influencing DB, a conceptual framework is outlined whereby DB is viewed in terms of different dimensions established within the Driver–Vehicle–Environment (DVE) system. Moreover, DB assessment has been approached by various machine learning (ML) models. Still, there has been no attempt to analyze the empirical evidence on ML models in a systematic way, furthermore, ML based DB models often face problems and raise questions that must be resolved. This article presents a systematic literature review (SLR) of the DB investigation concept; In the first phase, a framework for conceptualizing a holistic approach of the different facets in DB analysis is presented, as well as a scheme to guide the future development and implementation of DB assessment strategies. In the second phase, an overview of the literature on ML is designed, revealing a premier and unbiased survey of the existing empirical research of ML techniques that have been applied to DB analysis. The results of this study identify an interpretive framework incorporating multiple dimensions influencing the driver's conduct, in an attempt to achieve a thorough understanding of the DB concept within the DVE system in which the drivers operate. Additionally, 82 primary studies published during the last decade and eight broadly used ML models were identified. The findings of this review prove the performance capability of the ML techniques for assessing DB. The models using the ML techniques outperform other conventional approaches. However, the application of ML models in DB analysis is still limited and more effort is needed to obtain well-formed and generalizable results. To this end, and based on the outcomes obtained in this work, future guidelines have been provided to practitioners and researchers to grasp the major contributions and challenges in the state-of-the-art research. Highlights: Driving behavior is a complex concept defining the driver's different states. Conceptualizing a framework for driving behavior within the surroundings. A systematic review of Machine Learning models for driving behavior analysis. Performance of Machine Learning models foster driving behavior assessment. Machine Learning models outperform other conventional approaches.