Pedestrian safety is one of the most discussed topics in vehicle safety right now. There are concerns regarding the realistic implementation of real world accident issues into the EEVC WG17 component testing procedure. The European car industry presented a modified component test in relation to the self commitment. The key advantages of the current test procedure are easy handling and high reproducibility. They are contrasted by a number of specific problems which the current component test can not address. For example it can not reproduce the effect of the shape of the car front on the kinematics of a colliding pedestrian. The contact points of a pedestrian on a car front are car-specific. Therefore no predetermined test zones should be used. For the determination of these car-specific test-zones numerical simulations may be used. The presented approach for a test procedure combines numerical simulations and component tests into a hybrid-test; it is able to solve most of the mentioned disadvantages of a conventional component test without complicating it unduly. The numerical simulation allows to define the car-specific parameters of the pedestrian-car-collision in terms of localization of the contact, impact angle and velocity of the relevant pedestrian body parts. These parameters are in a second step used as input for the experimental component test. This hybrid-test reproduces real world pedestrian-car-collisions much better. It can be applied to all current and future car concepts (sport utility vehicle, minivans, crossover concepts etc.) very easily. This method can also be used in the early stages of car development, to improve the preconditions for pedestrian safety. This results in better cost effectiveness. The key ideas of the hybrid-test will be presented in the paper: Starting with a description of the pedestrian-car collision a suitable numerical model was created. Multibody dummies are used to collide with passenger cars under a variety of conditions (size of the pedestrian, relative location of car and pedestrian, relative speed). The TNO-pedestrian model has been chosen. The procedure has been applied to two very distinct car models. As a result a statistical pattern describing the impact of pedestrians in a collision is generated for the two selected cars. It is shown that the results are considerably at variance to the testing conditions according to EEVCWG17. For the covering abstract see ITRD E825082.
Abstract