Think Aloud Protocol and Decision Tree for Driver Behavior Modeling at Roundabouts

Learning from human driver’s strategies for undertaking complex traffic scenarios has the potential to improve decision-making methods designing ADAS systems, as well design self-driving rules automated vehicles. This paper proposes a human-like algorithm built up drivers experiential naturalistic driving. The approach of this work consists exploring two main techniques. Firstly, use “think aloud protocol” build dataset based on driving, capturing intentions. Afterwards, technique decision tree is used generate an categorize driving patterns function circumstantial parameters. study focused simple roundabouts in presence other translated into algorithmic rules, where pathways are represented ’if-then’ rules. Finally, accuracy driver behavior model been assessed, yielding promising results.