Comparison Of The Virthuman Model And Hybrid Iii Dummy At Low-Speed Frontal Impact

Abstract

"Current trends in mobility bring new challenges for both active and passive safety. Although connected and highly automated vehicles shall be impact-free due to the common communication, safe co-existence between automated and non-automated conventional traffic for a long transition period of mixed traffic must be ensured. Current approaches for vehicle certification process involves the mechanical dummies, which are usually limited to a single direction impact assessment. Virtual human body models bridge the gap enabling assessment in multi-directional impact scenarios. Different human body models have been already implemented for safety assessment in high-speed impacts, however, low-speed impact scenarios are also necessary to be addressed. The main aim of this work is to show the biofidelity of the virtual human body model Virthuman in the low-speed crash scenario. The exploited hybrid scalable virtual human body model Virthuman is formed by the skeleton as a multi-body structure coupled to deformable segments representing the outer skin. Thus, the model can be simply adapted to any initial position in the virtual environment for the fast calculation process for injury risk prediction. The model is scalable, so it is simply able to represent a human subject of the given age and posture. The special sled test device is used in the experimental measurement with the human volunteer and four-times with the Hybrid III dummy to simulate the low-speed impact. The low-speed impact scenario is numerically simulated using the finite element model of the Hybrid III dummy as well as with the hybrid Virthuman model. The simulation with the finite element Hybrid III is firstly used for the proper definition and validation of the seat belt model, which will be further used also in the simulation with the Virthuman model. The biodelity of the Virthuman model compared to the biodelity of the physical Hybrid III dummy is tested via objective correlation and analysis (CORA) method. The results show that the human body model performs better real human body behaviour than the dummy model in the low-speed impact. The presented work summarizes a possible approach towards virtual prototyping of safe interior concepts in the automated driving era. Acknowledgement: The work was supported by the European Regional Development Fund-Project Application of Modern Technologies in Medicine and Industry” (No. CZ.02.1.01/0.0/0.0/17_048/0007280) and by the internal research grant SGS-2019-002. Special thanks belong to BESIP providing the low-speed sled test environment and ŠKODA Auto a.s. for providing the infrastructure and the dummy."