The risks incurred by the head of restrained front occupants in real frontal impacts must be correctly reproduced in experimental tests (especially statutory tests) to be sure that improvements made will be effective in real accidents.
A comparison of the frequency and severity of head impacts for 179 real accidents and 211 impact test dummies is performed over a range of velocity changes between 46 and 60 km/h.
The results show major differences between the risks incuned in real accidents and the risks predicted by global frontal impact against a 0° angled barrier, applying inadequate and outdated criteria.
Possible improvements are suggested.