EVALUATING DEGREE OF DANGER IN DRIVING BEHAVIOR WITHOUT RIGHT OF WAY WHEN ENTERING A NON-SIGNALIZED INTERSECTION

摘要

For fundamental education about careful driving to prevent right-angle collisions when entering a non-signalized intersection, it is important for drivers without the right of way to understand the problems with their driving behavior and how they should improve it. In our previous study, we analyzed actual driving behaviors of some drivers recorded on the road. As a result, we supposed that drivers who exhibit dangerous driving behaviors could not make a sufficient safety check for potential hazards that they didn't actually see. So we constructed a simulation system that evaluates the potential risk for each driver using a simulation program. The driver model used in the simulation program was based on five kinds of elemental driving behaviors that characterize a scries of driving behaviors when entering a non-signalized intersection. In other words, the simulation system kept driving behaviors in perspective and indicated in particular the main point that drivers should improve. However, a more detailed difference in driving behaviors when approaching the stop-line upon entering an intersection would strictly determine whether a driver will get into a collision or not. So it is essential for fundamental education about careful driving to evaluate driving behaviors in detail and to teach the points that a driver should improve according to his/her characteristics and the degree of his/her dangerous driving behaviors. In this study, we constructed a system that evaluates driving behaviors in detail by using a series of data. In evaluating driving behaviors, car's speed and driver's eye movement will be very significant factors. However, from the viewpoint of careful driving, these driving behaviors are not significant independently. They should be evaluated together as a synthetic index. When entering a non-signalized intersection, a driver maintains safety for moving ahead by making a visual safety check. Continuing such observation enables a driver to pass an intersection safely. From this viewpoint, we attempted to calculate the potential risk level toward imaginary crossing cars and bicycles by using data on a car's position, its speed and a driver's eye movement at a certain time. In this calculation, the area from which crossing cars or bicycles could appear was determined by considering a driver's eye movement that he/she had carried out previously. So this calculation can evaluate the risk revel of driving behavior include the importance of making safety checks. Next, we made a simulation program that expresses the risk level of driving behaviors in time-scries. This simulation program shows graphs that express the risk level toward four kinds of objects, crossing cars and bicycles from the right and left sides. By using this simulation program, a driver who exhibits dangerous driving behaviors will visually understand what is dangerous in his/her driving behaviors. Finally, we evaluated the risk level of some drivers by using data recorded in actual driving on the road. We also examined the peculiar problem of each driver's driving behaviors. In the future, if we can record driving behaviors in real time, this method will be useful for concrete instruction in a driving support system.