When a pedestrian is thrown through the air by the impact of a motor vehicle, the distance travelled by the pedestrian from the point of impact can be used to provide an estimate of the speed at which the vehicle was travelling. Three sets of data collected from real accidents are reviewed with the aims of assessing the value of this method and fixing the model used to estimate the speed. The data have been published by Stuertz and Suren, by Hill and by Dettinger. Each provides the vehicle speed estimated from skid marks and the pedestrian throw distance for a number of real traffic accidents. The parameters in a simple, physically motivated model are fitted to each dataset separately using a least squares method. It is found that the earlier dataset has a much less consistent relationship between throw distance and vehicle speed than the later sets. The two more recent sets of data are used to construct a model that provides an estimate of the vehicle speed based on the throw distance, the bounds within which, with 95 per cent certainty, the vehicle speed lies, and an absolute lower bound. The data are also used to assess a method considered by Wood for the estimation of vehicle speed from the distance between the final resting place of vehicle and pedestrian. It is concluded that this last method is not likely to be useful because of its sensitivity to errors in measurement. (Author/publisher)
Samenvatting