Visual Search Efficiency Evaluation Method for Potential Connected Vehicles on Sharp Curves

Abstract

Accurate discrimination of driving proficiency is essential to improve driving safety. A naturalistic driving experiment was designed and organized to discover the response characteristics among drivers’ eye movement behaviors, driving proficiency, and mountain roads; 20 drivers were recruited to collect corresponding data of eye movement parameters on sharp turns, including gaze behavior and saccade behavior. The data show that on right turning roads, the main range of eye fixation points is −0.1–0.3 m in the horizontal direction, while the range is −0.35–0.2 m for left turn bends. The main range of the eye horizontal gaze angle is 0–30°, the range of the saccade amplitude is 5–25°, and its value increased as the curve radius decreased, showing a logarithmic relationship. Based on the collected eye movement characteristics data, a driver’s visual search modal matrix was built. A principal component analysis was used, and a comprehensive evaluation model of drivers’ eye movement characteristics was established. The result shows that eye movement comprehensive score and driving mileage were positively correlated. The result also indicates that the discrimination model can be used to quantify and discriminate driving proficiency and can provide a novel perspective for connected vehicle performance assessment.