Driver Classification to Clarify Behavior of Elderly Drivers

Abstract

The study aims to calssify older drivers in order to clarify their driving behavior. These classification approach will help us to understand their characteristics as well as to develop driver assistance systems and education programs in order to prevent trafic accidents. We focused on three aspects of drivers, collision avoidance, general driving and driving style. We first applied cluster analysis to driver parameters gained by driving simulator and questionnaire studies. The result classified subjects into several clusters. We then carried out linear discriminant analysis to describe linear discriminant functions. Using the fucntions, we can classify them with high probability. Evaluation of functions also showed characteristics of each cluster.