Technical Sessions

Dr. Nobuhisa Tanaka, National Traffic Safety and Environment Laboratory, Japan
Dr. Kazumoto Morita, National Traffic Safety and Environment Laboratory, Japan
Dr. Michiaki Sekine, National Traffic Safety and Environment Laboratory, Japan
Dr. Yuki Tsukada, National Traffic Safety and Environment Laboratory, Japan
Mr. Shunsuke Takagi, National Traffic Safety and Environment Laboratory, Japan
Dr. Toshiya Hirose, National Traffic Safety and Environment Laboratory, Japan
Dr. Yoshiroh Aoki, National Traffic Safety and Environment Laboratory, Japan
Dr. Tomonori Hasegawa, National Traffic Safety and Environment Laboratory, Japan
Mr. Kazuo Matsushima, National Traffic Safety and Environment Laboratory, Japan

The Japanese government is aiming to decrease annual automobile traffic accident fatalities to below 5,500, and to reduce the annual number of deaths and injuries from traffic accidents to less than one million, by 2010. To achieve these targets it will be necessary to implement effective vehicle safety measures, and the early dissemination of effectual active safety support systems on the market is desired. This will require objective evaluation of the effect that various safety systems have on reducing accidents. The authors are currently developing a simulation program (simulator) that can evaluate safety support systems. This simulator virtually runs vehicles equipped with a given safety system and reproduces traffic conditions on a computer. It generates traffic flow over for long term using Monte Carlo simulation, and compares the number of accidents and near misses that occur in vehicles with and without the safety system to estimate the effect of the system. This method, based on virtual running of a vehicle with the safety system being evaluated, estimates the possibility that an accident will occur. By operating virtually, it has the characteristic that it can evaluate even safety systems which have not spread on the market. In addition, since driver characteristics may also change, it can estimate the incidence of accidents for a group of drivers that have a given characteristic. Real-world accidents involve complex relations between driver, automobile, and road and environmental factors, and it is necessary to simulate those factors appropriately. We first attempted to simulate intersection collisions, which are responsible for many deaths and injuries, and treated Brake Assist Systems (BAS) in relation to safety systems. Therefore, we assume a vehicle with degrees of freedom in the front-back direction, on the rotation of the front wheels, and on the rotation of the rear wheels. For other safety systems, we made basic program settings that can be evaluated in the future. Driver factors are closely related to the occurrence of traffic accidents, and accidents need to be understood on the assumption that humans make errors. The average driver makes repeated perceptions, judgments, and physical manipulations when operating a vehicle. Thus, the driver model is one consisting of three levels-a perception level, a judgment level, and a manipulation level-with human error occurring at each one. We have developed a simulator that realizes these elements, and we are currently analyzing this simulator.

Session: Advanced Safety Systems I

• Print version
• Email this page