摘要 |
PROBLEM TO BE SOLVED: To control the constraining of an occupant constraining means based on a detection result by detecting whether or not an occupant of a vehicle recognizes a possibility of a collision before the collision of the vehicle. SOLUTION: The face of an occupant 1 is photographed by a camera 13. A light of sight detection circuit 14 detects a line of sight direction of the occupant 1 by using an image processing from the photographed face image and successively stores the results. Next, a collision recognition detection circuit 15 investigates time series data of the light of sight directions for the past 10 seconds at a point of time. Assuming that a front collision occurs just after the time, whether the occupant 1 recognizes the collision or not is judged by depending on whether the occupant 1 looks at the front for the 10 seconds. In this case, the occupant 1 is considered to recognize the collision if the occupant 1 looks at the front even once for the 10 seconds. When the occupant 1 is judged that he/she not recognize the collision, a control method setting circuit 16 increases gas quantity of an air bag 2 more than that at the time of a normal collision and further sets tensile force of a seat belt 3 slightly weaker than that at the time of the normal collision. |