The pedestrian model in PC-Crash is based on a multi-body system, where several bodies are interconnected by joints. Each of these bodies can have different properties to represent the different parts of the human body. The joint properties can be specified independently as well. The theoretical background of the pedestrian model has been introduced in SAE 1999-01-0445 and the model shows to give a good correlation of the gross movement of the pedestrian compared to crash test data.
As there are many parameters, which can and have to be specified for the pedestrian model as input parameters, an in depth validation of the different parameters has to be done to validate this model.
This paper describes in detail the validation process for the pedestrian model. A significant number of crash tests (approx. 30) was used as a basis to compare the results of the simulations and the real movement of the test subjects. Different test configurations were selected to provide a set of crash tests, which represents real cases as close as possible. For this reason crash tests with different vehicle types and vehicle shapes were investigated and crash tests with different driving conditions (braking, sliding) were used as well. To show the difference between dummy movement and the movement of human test subjects some tests with human test subjects and some real cases, which were very well documented, have been used. Each of the different tests was reconstructed using PC-Crash and the pedestrian model and the significant difference between simulation results and crash tests are pointed out. A special interest has been focused on the pedestrian trajectories and the contact locations between the pedestrian and other objects.
All tests and simulations are compared in the conclusion and the usability of the pedestrian model in different accident configuration is discussed. Typical parameter ranges for the different input parameters like pedestrian geometry, weight and stiffness parameters are presented as well.