A Decision-making Model for Explaining Driver Behaviour

MSc. Marjana Čubranić-Dobrodolac

Doctoral thesis

Abstract:

The main topic of this dissertation is the modeling of driver behavior based on an examination of their psychological traits. After a detailed review of relevant literature, five questionnaires have been prepared to collect the data. Four questionnaires are related to testing the psychological constructs of drivers and an additional one is a demographical and driving history questionnaire. A survey was carried out at the sample of 305 drivers, from which there were 202 professional drivers and 103 drivers of privately owned vehicles. The data were processed by two general approaches: statistical and fuzzy logic. The implemented statistical methods are hierarchical regression analysis and binary logistic regression. The driver behavior is modeled by fuzzy inference systems where the inputs are the results from psychological tests and the output is the number of experienced road traffic accidents in driving history. The performance of a fuzzy inference system that can be considered as a decision?]making tool for explaining driver behavior, is further enhanced, in the sense of adjusting its results to the empirical data, by applying the bee colony optimization metaheuristic. Based on the obtained results, adequate recommendations for traffic safety improvement are proposed.

Abstract:

The main topic of this dissertation is the modeling of driver behavior based on an examination of their psychological traits. After a detailed review of relevant literature, five questionnaires have been prepared to collect the data. Four questionnaires are related to testing the psychological constructs of drivers and an additional one is a demographical and driving history questionnaire. A survey was carried out at the sample of 305 drivers, from which there were 202 professional drivers and 103 drivers of privately owned vehicles. The data were processed by two general approaches: statistical and fuzzy logic. The implemented statistical methods are hierarchical regression analysis and binary logistic regression. The driver behavior is modeled by fuzzy inference systems where the inputs are the results from psychological tests and the output is the number of experienced road traffic accidents in driving history. The performance of a fuzzy inference system that can be considered as a decision?]making tool for explaining driver behavior, is further enhanced, in the sense of adjusting its results to the empirical data, by applying the bee colony optimization metaheuristic. Based on the obtained results, adequate recommendations for traffic safety improvement are proposed.

Keywords

psychological traits, driver behavior, traffic accidents, hierarchical regression analysis, binary logistic regression, fuzzy inference system, bee colony optimization

Language used: Czech

Date on which the thesis was submitted / produced: 7. 1. 2021

Thesis defence

• Date of defence: 12. 3. 2021
• Supervisor: doc. Ing. Libor Švadlenka, Ph.D.

University of Pardubice

Faculty of Transport Engineering

Doctoral programme / field:
Technique and Technology in Transport and Communications / Technology and Management in Transport and Telecommunications: Management