Applications and limitations of wrap-around ratio to vehicle speed estimation in pedestrian collision analysis

Abstract

Road traffic injuries could become the fifth leading cause of death globally by 2030 unless appropriate countermeasures are taken. Reliable reconstruction of collisions is a prerequisite for this, and determination of vehicle impact speed is the critical reconstruction parameter for vehicle-pedestrian collisions. In this paper, a validated Constant Inertial Property pedestrian model is applied in a Monte Carlo simulation environment to assess the feasibility of using pedestrian wrap-around ratio in the estimation of vehicle impact speed for collision scenarios involving a range of pedestrian pre-impact gait positions, anthropometrics, vehicle shapes and braking conditions for wrap type pedestrian impacts. The results show that there is a clearly identifiable increase in wrap-around ratio with vehicle impact speed up to a speed in the region of 40 km/h, after which further increases in impact speed produce very little change in wrap-around ratio. This characteristic is confirmed from fundamental impact and kinematic analysis. Therefore, the results show that for higher-speed cases only minimum speed predictions may generally be reliable. Statistical analyses of the Monte Carlo simulation results are then used to provide Confidence Limits and Odds Tables for accident reconstruction purposes. These provide the first quantitative approach to vehicle speed prediction from pedestrian wrap-around ratio.

Predicting the pedestrian pre-impact speed from the pedestrian projection distance and vehicle damage measurements

“…Using the linear relationship between the ratio V p /V c of the pedestrian speed to the vehicle speed and the ratio H/L of the transverse offset to the longitudinal offset, the pedestrian speed can be estimated, given that H/L can be measured from vehicle damage patterns and that the vehicle speed can sometimes be determined from the pedestrian projection distance [8][9][10][11] or the wrap-around distance. 23 However, the transverse offset and the amount of head rotation about the coronal plane (see Figure 2) between the primary and secondary pedestrian contacts with the vehicle vary with the pedestrian gait in a cyclical manner, which has been previously modelled using third-order Fourier curves. 13 Monte Carlo simulation based on this modelling yields the distribution of head rotations u as a function of V c from data set 2 (see Figure 12), which shows that the density of head rotation angles is highest towards the positive and negative extremities of u angles.…”

“…12 It has also been shown to predict the pedestrian wrap-around ratio over a broad range of vehicle impact speeds. 23…”

“…As previously outlined for vehicle speed predictions, 8,10,23 speed estimates in pedestrian-vehicle collisions are required for various purposes, including general traffic safety studies, biomechanical injury research and litigation. A degree of uncertainty is inevitable and therefore a confidence interval is more meaningful than a single speed estimate, but the acceptable range of uncertainty depends on the proposed use for the prediction.…”

“…The influence of the vehicle front geometry (dive or no dive) on the longitudinal damage offset L has been reported elsewhere. 21,23 In the presence of vehicle dive, the longitudinal distance L is increased primarily owing to the changed vehicle front geometry. For braking without dive the only factor influencing the longitudinal distance L is the change in the car speed as a result of braking during the time lapse between the front and head impacts.…”

“…On this basis, the data presented in this paper can be applied for the two cases of no vehicle braking and vehicle braking without dive (initial stages of braking) for collision speeds above 40 km/h. The distribution of vehicle braking and vehicle dive was interpreted from a review of accidents between 2002 and 2009 31 and accidents reported by Wood et al 23 These reports showed that the distribution of pedestrian accidents is as follows: no braking, 43%; braking without dive, 33%; braking with dive, 24%. On this basis, the model would be applicable in about 76% of cases.…”