Plausibility Assessment of Numerical Cyclist to Vehicle Collision Simulations based on Accident Data

According to recent traffic accidents statistics, the number of bicycle accidents has not been decreasing since 2010 and is even increasing in Germany. In the past, human body models were used to gain further insight into the kinematics and injury prediction for pedestrians and were validated for vehicle collisions based on post mortem human subject test data. As comparable validation data is not available for cyclists, data from accident databases offers the opportunity to compare simulation results with reality. In this study, three accidents were selected where the dimensions of the involved cyclist and vehicle matched those of the selected human body model (THUMS) and of the car model (Toyota Camry). The boundary conditions of the collision simulation were defined according to the GIDAS documentation. Data from the simulation on collision points, injury-inducing contact pairs and injury predictions were compared with GIDAS. The results of the simulations mostly matched the documented collision data very well. The data suggests that finite element simulation using state-of-the-art human body models can be used to reconstruct accident data, thus increasing trust in the simulation of collision scenarios, where a validation is currently not possible.