Comparison of multibody and finite element human body models in pedestrian accidents with the focus on head kinematics

<p><b>Objective:</b> The objective of this study was to compare and evaluate the difference in head kinematics between the TNO and THUMS models in pedestrian accident situations.</p> <p><b>Methods:</b> The TNO pedestrian model (version 7.4.2) and the THUMS pedestrian model (version 1.4) were compared in one experiment setup and 14 different accident scenarios where the vehicle velocity, leg posture, pedestrian velocity, and pedestrian's initial orientation were altered. In all simulations, the pedestrian model was impacted by a sedan. The head trajectory, head rotation, and head impact velocity were compared, as was the trend when various different parameters were altered.</p> <p><b>Results:</b> The multibody model had a larger head wrap-around distance for all accident scenarios. The maximum differences of the head's center of gravity between the models in the global <i>x</i>-, <i>y</i>-, and <i>z</i>-directions at impact were 13.9, 5.8, and 5.6 cm, respectively. The maximum difference between the models in head rotation around the head's inferior–superior axis at head impact was 36°. The head impact velocity differed up to 2.4 m/s between the models. The 2 models showed similar trends for the head trajectory when the various parameters were altered.</p> <p><b>Conclusions:</b> There are differences in kinematics between the THUMS and TNO pedestrian models. However, these model differences are of the same magnitude as those induced by other uncertainties in the accident reconstructions, such as initial leg posture and pedestrian velocity.</p>