A STUDY OF PERSONAL ADAPTIVE DRIVING SUPPORT SYSTEM USING A DRIVING SIMULATOR

This paper describes how the social damage by a traffic accident is very enormous, therefore various countermeasures are taken for accident reduction. Most of accidents are caused by a human-error. Some researches and developments of the driving support system are executed to prevent this human-error. These systems early detects an error of a driver in "recognition" "judgment" "operation", then a sound and visual information are shown to the driver for the accident avoidance. This driving support system controls the vehicle by steering and braking when the driver can not avoid a collision. A driving simulator (DS) is an effective tool for the development of the driving support system and the evaluation of HMI (Human Machine Interface). The authors have developed a Toyota Research Driving Simulator (TRDS) to study the Personal Adaptive Driving Support System (PADSS). PADSS is able to detect an omission or deviation from normal operation by comparison the difference between current driving operations and beforehand accumulated one. On the basis of this detected result, PADSS assists the safety driving by the warning and vehicle control when a driver falls into dangerous situation. This paper describes TRDS and the analysis result of accumulated driving characteristics. Furthermore, some results of the PADSS are shown. TRDS is constructed from the 6 Degrees of Freedom (DOF) motion platform and the longitudinal motion device. The 6 DOF motion is constructed by link mechanism, which mounted on the longitudinal motion base. The range of longitudinal motion is ±4 m. The visual system is reflected with a liquid crystal projector from a viewpoint of a driver to the three flat screen installed in front 1.4m. Field of View (FOV) of TRDS is 150 degrees. The driving characteristics are accumulated while a subject drives to a destination of the neighbor town via a highway from a certain departure place. Subjects including two women are 12, and the trial number is 20 times. The characteristics of each subject in braking are automatically extracted during the experiment, and new data are added to a database after the experiment. On the basis of braking characteristic of the accumulated individual subject, the authors carried out experiment of PADSS to show a warning for rear-end collision accidents. As a result of experiment, it was understood that the warning frequency is quite difference between each subject. There were many subjects who evaluated as the unnecessary about the rear-end collision warning. However, they had done hardy braking operation with the case which evaluated necessity of a warning. It is a future problem to decrease this unnecessary warning.