To develop a human-friendly driver support system, it is essential to detect the driver's status such as consciousness levels and looking aside. We have developed a driver monitoring system, which detect the driver's consciousness reduction and gaze direction change by image processing techniques. Furthermore, we have proposed a new warning method with this driver monitoring system. The effectiveness in this method has been verified using a driving simulator.

No abstract

SUMMARYThe driving simulators hitherto developed were intended for investigation of active safety techniques, verification of vehicle movement models, etc., without using an actual vehicle. But there subsequently arose a need for a system that could be used to investigate ways of providing driving and safety assistance to drivers in the aging society of the future. There has been no driving simulator available for such evaluations. The authors have developed a driving simulator for evaluation of driver-assistance systems with the following installations. A driver monitor camera is installed in the inner mirror unit of the car, and the level of consciousness of the driver is estimated by video processing. A technique for measuring the gaze direction is installed. Techniques for character size transformation in the display and warnings are investigated with reference to the lowered visual acuity of elderly drivers. A verification experiment is performed on braking distance, which is affected by the increased reaction time when the level of consciousness is low. The effectiveness of the system as verified by experiment is reported.

No abstract