Abstract

Low-speed rear-end vehicle collisions account for a considerable proportion of the road safety burden. A particular challenge for legal proceedings is the great difficulty in establishing the veracity of insurance claims, and the presence or absence of vehicle damage is often the only objective data available to assess collision severity. In this paper we combine a linear spring model with literature-based injury reference levels and vehicle damage regression models to provide novel insights into injury/symptom reference level probability in low speed rear-end collisions based on observed damage patterns. Analysis of staged collisions performed by the accident mechanics group AGU and others was first used to confirm that the impact behaviour of vehicles in low speed rear-end collisions is well represented by an equivalent linear-spring model. Accordingly, we represent a single overall normalised stiffness distribution for low speed rear-end collisions. Variation in this relationship showed the importance of vehicle bumper stiffness in the resulting vehicle acceleration patterns for a given collision speed. Further, integration of this model with acceleration-based lower and upper injury/symptom reference levels from the literature showed the dominant effect of vehicle stiffness on injury/symptom likelihood for a given collision speed. Application of empirical damage assessments from staged tests showed that a low speed rear-end collision cannot simply be judged as being of ‘minimal impact’ from observation of damage alone. The combination of the linear stiffness model with the lower and higher injury/symptom reference levels and the damage probability characteristic can be applied to determine the overall proportions of collisions of any particular type which exceed injury/symptom reference levels. We do caution that the injury/symptom reference levels employed here are essentially exemplar, with the occurrence of injuries in specific cases depending also on other factors.