Research On Occupant Collision Injury Risk Prediction For Intelligent Vehicle Safety Strategy

Road traffic safety and human casualties in global traffic accidents have always been a major problem worldwide.The development of intelligent vehicle technology has brought new opportunities and challenges to road traffic safety.This research takes the protection of intelligent vehicle as the background,and estimates the injury risk faced by the occupant in different scenarios based on the traffic environment and occupant status information perceived by the intelligent vehicle,and then the injury risk of occupant is reduced by emergency intervention in the vehicle state and changing the configuration of the safety system.In this research,a human body model with high degree of biological fidelity was selected as the main analysis tool.The influence of traffic accident parameters on human injury results was studied through numerical simulation,and the injury prediction models for different scenarios were generated.Finally,an injury minimization-oriented intelligent vehicle safety protection strategy is established,with injury prediction models as the core.First of all,in order to explore the influence of human body differences on injury results in traffic accidents,a set of finite element human body model parameterization process is established for the benchmark human body model,which could be used for scaling of baseline human body models.Based on the Chinese anthropometric database,human body models capable of representing Asian population stature were developed.The developed models have better mesh quality and more accurate anthropometric parameter values than existing models.Then,taking lumbar spine and brain as examples,the potential problems of usage of some existing injury risk criteria in vehicle safety analysis are discussed.Through a typical traffic accident reconstruction,the three-stage loading process of the lumbar spine in a frontal crash was reproduced,and it was found that the complex loading pattern and process would significantly reduce the injury threshold of the lumbar spine.Based on experiment reconstruction,it was found that the strong connection between the skull and the brain would increase the stress out put of the model and reduce the strain output,and the mechanical loading mechanism of the brain during the whole process had two forms of pulling and slapping.Next,the dangerous state of the vehicle nearing a collision is defined and injury prediction models are developed for a simple frontal collision condition.It was found that the differences in the form of prediction models would affect the complexity of the information contained in the results.A typical two-vehicle collision condition was extracted from the traffic accident database,and a four-parameter two-vehicle collision condition characterization method was proposed.Asian population was focused as the occupant object to establish the injury prediction model.Through the establishment of the injury risk safety domain,it was found that the average injury probability of the occupants of MAIS 3+differed by 10% when the two vehicles crashed at different impact positions and angles.Finally,with the injury prediction models as the core,an inju ry minimization-oriented intelligent vehicle safety strategy is established b ased on the two conditions of frontal collision and two-vehicle typical collision.Through example analysis,it is found that the safety strategy based on injury optimization can guide the vehicle manipulation to the condition with the lowest risk of occupant injury,and a safety strategy that conforms to intuitive perception is obtained in the simple frontal collision condition.