Predicting head impact conditions in vehicle-to-pedestrian impacts through computational human modeling

This study aimed to use finite element (FE) pedestrian and vehicle models to generate a virtual database of pedestrian impacts and develop prediction models for pedestrian head impact conditions, which are important to evaluate the effects of vehicle front-end designs on pedestrian head injury responses. The generic vehicle (GV) models used in Euro NCAP originally developed based on European vehicles were morphed into 20 U.S. vehicle front-end geometries across a wide range of vehicle types and characteristics. A total of 240 FE pedestrian impact simulations were conducted using the 20 morphed GV models with four sizes of pedestrian human body models (6-year-old, small female, midsize male, and large male) at three impact speeds (30, 40, and 50 kph). A set of predictors, including vehicle front-end geometry, pedestrian size, vehicle impact speed, and pedestrian wrap around distance (WAD) were selected based on literature to predict head impact time (HIT), head contact velocity, and head contact angle. R2 values and root-mean-square-error (RMSE) were used to evaluate the quality of the prediction models. High correlations and good accuracies were achieved in the prediction models for HIT (R2 = 0.979 and RMSE = 6.61 ms), head impact velocity (R2 = 0.799 and RMSE = 1.39 m/s), and head impact angle (R2 = 0.846 and RMSE = 7.95 deg for adult pedestrians). It was found that impact speed, WAD, hood angle, and hood height are statistically significant variables for predicting the pedestrian head impact conditions. HIT is highly predictable in pedestrian impacts, while the head impact velocity and head impact angle are associated with larger variations in the selected impact conditions. This indicates a potential need of varying impact velocity and angle for future vehicle evaluations of pedestrian head protection. This study generated a virtual database of pedestrian impacts with a wide range of vehicle front-end geometries, and developed prediction models to use vehicle front-end geometry, pedestrian size, impact speed, and WAD to predict pedestrian HIT, head contact velocity, and head contact angle.