Classification of automated vehicle’s driving style by adapting multidimensional driving style inventory for Malaysian driver

• Auteur principal: Kamaludin, Muhammad Zainul Abidin
• Format: Thèse
• Langue: anglais; anglais
• Publié: 2024
• Accès en ligne: http://eprints.utem.edu.my/id/eprint/28843/; https://plh.utem.edu.my/cgi-bin/koha/opac-detail.pl?biblionumber=124493

The rapid advancements in automated vehicle technology have sparked questions about how these vehicles will behave on the road. To address this, must first understand the driving styles of human drivers and then predict the driving style of automated vehicles. The Multidimensional Driving Style Inventory (MDSI) is a commonly used tool to categorize human driving styles in different cultural contexts. This study's goal is to adapt the MDSI for use with Malaysian drivers and determine the specific driving profiles that exist among them. Additionally, the study examines the relationship between driving style and personality traits, as well as sociodemographic information. A total of 737 drivers, ranging in age from 17 to 49 years, completed the MDSI and a questionnaire on personality traits (trust, desire for control, and sensation seeking). Through Exploratory Factor Analysis (EFA) and Confirmatory Factor Analysis (CFA), the study found established a 5-factor structure for the MDSI, comprising risky, anxious, angry, careful, and dissociative driving styles. CFA confirmed that the model fit of the MDSI was acceptable. Further analysis revealed that some respondents exhibited only one driving profile, while others displayed multiple profiles. The study found that the MDSI factors had moderate correlations with "desire for control" but weaker associations with "trust" and "sensation seeking." Significant differences in driving styles were observed based on gender, age, experience, and the type of car driven. The study then proceeded to its second phase, which aimed to validate the five driving style factors (careful, risky, angry, anxious, and dissociative) among Malaysian drivers through on-road observational studies. Participants were asked to engage in driving experiments using instrumented vehicles on two designated routes, with 45 minutes of recorded driving tasks. The results indicated a modest correlation between the MDSI scores and the data obtained from on-road observations, particularly regarding acceleration in the x, y, and z directions. The findings from both phases of this study demonstrate that Malaysian drivers exhibit distinct driving profiles consisting of these five factors, and these profiles correlate with acceleration data. This data is essential for developing models of driving styles and accelerating the acceptance of user-driven automated vehicles in the near future