| Two-wheelers are becoming an increasingly popular choice for travel because of their convenience and flexibility,and it is extremely common for manned two-wheeler,with accidents occurring frequently.In case of traffic accidents,cyclists and occupants are more likely to be injured due to the lack of protection from body shells and airbags.In this paper,we use real two-wheeled vehicle accidents as a research object to explore the reconstruction methods of two-wheeled vehicle accidents and the differences in injuries and kinematic responses of cyclists and occupants.This paper is based on 170 videos of real car-manned two-wheeler accidents,obtains information on weather conditions,collision patterns,collision speeds,and driving status in accidents,and summarizes the general characteristics of the occurrence of manned twowheeler accidents after statistical analysis.The results showed that the accidents occurred mostly at intersections or one-way streets;the most common was the collision between cars and electric two-wheelers or ordinary cross-type motorcycles;the speed of cars was concentrated between 10~50km/h and the speed of two-wheelers was concentrated between 10~40km/h;the most common collision pattern of two vehicles was side right-angle collision.Then,the typical collision scenes of manned two-wheeler accidents were extracted by using cluster analysis method with parameters of vehicle speed,two-wheeler speed,twowheeler type,and driving status as variables to obtain eight types of typical scenes of manned two-wheeler accidents.In order to obtain the injury and kinematic response differences of the riders and occupants under different scenes,first build the simulation platform based on PC-Crash,and extracted the method of reconstructing manned two-wheeler accidents based on PC-Crash;then a real manned two-wheeler accident was taken as an example to demonstrate and verify the reliability of the results obtained in reproducing the accident of manned two-wheeler accidents based on PC-Crash;further,in order to improve the reliability of the reconstruction results,two methods of uncertainty analysis of accident reconstruction results under complex uncertainty conditions,namely entropy judgment method and FDER,are proposed to address the problem of multiple models obtaining multiple different accident reconstruction results in accident reconstruction.Through three numerical cases and two real accident cases,found that both proposed methods can fuse multiple model results,and effectively reduce the probability value of unreliable results in the process of fusion.Finally,one typical case was reconstructed for each scene,and based on this,384 groups of simulation tests were designed and conducted based on PC-Crash with car speed,twowheeler speed and collision position as test variables according to the characteristics of each scene and comparative analysis of the riders and occupants injury and kinematic response data,it was found that there were significant differences in rider injury and kinematic response in most scenarios. |