Abstract
The relationship between driving behavioral indexes and near-miss events was made clear for the purpose of
developing a method of quantifying safe driving assessment based on accident risk. This study focused on right-turn
accidents, which are the second most intersection accidents in Japan. Candidate driver behavioral indexes were
considered based on typical accident patterns. Using a database of everyday driving behavior, driver behavioral indexes
having a significant correlation with the near-miss event rate were identified. Tests were conducted to validate the
indexes using an augmented reality experimental vehicle (JARI-ARV) to simulate a right-turning vehicle approaching
an oncoming vehicle ahead.
