Child restraint systems (CRS) for cars are intended to protect children in the case of a car accident. Unfortunately their effectiveness is still too low: in the range 30–50 % when it would be expected to be much higher. The low effectiveness of child restraint systems can partly be explained for the youngest passengers by their greater cervical vulnerability and for the oldest (from 3 to 12 years old) by the morphological immaturity of the pelvis. However, tools available to evaluate the effectiveness of CRS are very poor, as well as knowledge on injury mechanisms and criteria.
The CREST project was created to develop the knowledge on child behaviour and tolerances, the final aim being to propose new test procedures for determining the effectiveness of CRS using instrumented child dummies. Eleven partners were involved, namely Fiat Auto-SpA (with Elasis), INRETS, PSA Peugeot Citroën, Renault, TNO Automotive, TUB, RICE, BAST, GDV, MUH, VTI. The method used in this project was to collect data from accident investigations and from reconstruction crash tests in order to determine the physical parameters (forces, accelerations and deformations on the child) which correspond to the various child injury mechanisms. Hence, limits should be prescribed under which injuries could be avoided.
This paper presents roughly the methods used for the achievement of this project and the main results. In particular, data from the 56 accident reconstructions are presented and injury criteria are evaluated against reconstruction results.