3304 Driving Feature Extraction from High and Low Skilled drivers in Curve Sections Based on Machine Learning

Abstract

In order to improve drivers' driving skill, this paper proposes a statistical method to extract the operation characteristics between high-skilled and low-skilled drivers from obtained data at driving simulator experiments in curve sections. A driving skill determining model is built using the extracted features so that the driving skills of an unknown run can be classified automatically. And based on the extracted result, we supplied advices to low-skilled drivers in experiments. The results are also descripting in the paper.