Predictive capabilities of the MADYMO multibody pedestrian model: Three-dimensional head translation and rotation, head impact time and head impact velocity

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 essential for understanding and hence reducing the injuries sustained by pedestrians. The significant influence of vehicle speed, pedestrian speed and pedestrian gait on pedestrian transverse motion and rotation about the longitudinal axis has been qualitatively noted in the literature, but there has been no quantitative approach to this problem. The MADYMO multibody pedestrian model is widely used for collision reconstruction, but its validation to date mostly remains limited to body segment trajectories in the vertical plane along the direction of vehicle travel. In this article, the MADYMO pedestrian model is compared to staged tests and a real collision in terms of head trajectory, longitudinal and transverse head translations relative to the primary contact location of the pedestrian on the vehicle, impact location on the head (and hence longitudinal rotation of the body), head impact time and head impact velocity. It is shown that the model can reproduce staged cadaver and dummy tests in terms of head trajectory (mostly within 10%), longitudinal head translation (within 17%), transverse translation (two cases: errors of 0% and 19%), impact location on the head (within 45° in the majority of cases), head impact time (mean difference of 8.7 ms) and head impact velocity (mean difference of 1.8 m/s). A sensitivity analysis showed that the model is largely unaffected by changes in vehicle stiffness and vehicle–pedestrian friction, while variations in the pedestrian stance and the height of the vehicle front have a considerable effect on the kinematics of a collision. This is the first time that the MADYMO pedestrian model has been evaluated with regard to transverse motion and longitudinal rotation of the pedestrian. The results presented here mean that the MADYMO model is appropriate for further analysis of the influence of gait on pedestrian post-impact kinematics.