Research And Application Of Vehicle-Pedestrian Accident Reconstruction

Vehicle-pedestrian collision has become a major type of traffic accident all over the world. As the powerful tool of accident investigation, computer simulated reconstruction has recently attracted more and more attention in the field of traffic safety. Also, simulation based reconstruction serves as an important source of information to understand injury mechanisms and evaluate pedestrian safety of vehicle design. Considering the existing research and practical demand, this paper presents some improvements in pedestrian human model and accident reconstruction method which play an important role in pedestrian accident reconstruction.As for pedestrian human model, a commercial TNO pedestrian human model is firstly improved in order to realize the application to flat front vehicle-pedestrian accident reconstruction. For example, a new contact simulation method is adopted to avoid the contact instability so that the improved dummy can be suitable for more server loading conditions. Then, this study presents a method for rapid estimating the anthropometry parameters of a given victim and develops the dummy scaling tool, Chinese Pedestrian Model Generator (CPMG). This method and tool are based on dummy scaling technology and new anthropometry results of Chinese population, which can avoid using Western human population in China. Additionally, coupled FE-MB modeling technology is introduced to pedestrian human modeling for the reconstruction based on pedestrian injury. FE model of these key body segments should be modularized and combined into the MB full-scale pedestrian model, which has the ability of detailed prediction of pedestrian injury. In order to illustrate this method, a FE lower limb is modeled and combined into the MB pedestrian model to perform the analysis of long bone fracture in real-world cases. The model is validated at component and full-scale levels by the simulation with the tests on cadaveric specimens.As for accident reconstruction method, a series of full-scale dummy impact simulation are conducted to evaluate the influence of vehicle front end geometry, pedestrian initial stance and anthropometry in order to provide the engineering experience for real-world accident reconstruction. Also, automatic optimization based on rest positions and pedestrian injury is presented for pedestrian accident reconstruction. This method uses Sequential Linear Programming (SLP) which can automatically vary pre-impact parameters to minimize the error between simulation and real data. Our optimization procedures are performed with MADYMO/Optimizer coupled with MADYMO multi-body module and existing pedestrian models.Finally, the above-mentioned methods and models are applied to a series of real-world pedestrian accidents, of which 6 accidents are detailedly described in this paper. Due to the limit of physical experiment, the results of these reconstructions are proved by the accident data for example autopsy reports. The purpose of this application is to evaluate the capability and necessity of the above improvements.