According to the driving conditions of highways, the division of accident responsibility is mainly based on the identification of accident responsibility. In the process of accident identification, there exist many subjective factors, which cannot ensure the accurate identification and prevention of accidents. In this context, the paper first deeply analyses the causal relationship of accident factors using the accident survey data. Then, combined with the accident causation theory and error analysis method, the negative binomial regression model of random parameters was applied to analyse the relationship between major accident factors, and provide reliable subjective data support for accurate division of accident responsibility. Meanwhile, based on the theory of kinematic mechanics and the data of collision residual traces, it adopts the computer simulation technology to reconstruct the accident of vehicle collision, inversely calculate the motion state of the vehicle before collision, and thus provide relatively reliable objective data support for accurately dividing accident responsibility. Finally, using the error analysis method proposed in this study, the preventive measures should be taken specifically for the main accident-inducing factors, which has certain guiding significance for reducing the occurrence frequency of serious accidents on highways.
Keywords: highway accident; causal mechanism; division of responsibility; credibility; error analysis; simulation technology