In this paper are analysed the results of ten accident reconstructions of 5 frontal actual car traffic accidents. These accidents involved 9 restrained occupants, and were reconstructed first with 50th percentile dummies and then with human cadavers.
The results of these reconstructions are analysed in order to compare the injuries sustained in real accident first, with the injury criteria values recorded on dummies, and second with injuries and injury criteria values found on cadavers.
The results show that an AIS 3 head injury in actual accident could correspond to a low value of HIC recorded on dummy, that there is a large scatter in chest injury criteria related to chest injuries, and that if we need a protection for almost all occupants in frontal impact, we have to choose chest injury criteria value lower than the proposed one.