Analysis Of Injury Risk Of Two-wheeler Cyclists Based On Bayesian Network Model

Among all kinds of road traffic crashes, such as cars and bicycles, electrictwo-wheelers and other two-wheelers of car, collision accident rate increased gradually.And for the absence of protection of car body, riders have been the vulnerable groups inthe road traffic, vulnerable and mortality rates are relatively high. Therefore to preventand reduce two-wheeler rider injury accident has become the key topics of scholarsand foreign experts. Currently, the automobile and two-wheeler riders’ accident injuryanalysis are mostly based on people, vehicles, roads and environment, a single factor,isolated analysis, which makes the research and analysis methods not universal.Based on the accidents causes analysis,the car and two-wheeler collision causesanalysis were combined with the Bayesian network approach, to carry out the researchof two-wheeler rider risk injury. Considering the complexity and diversity oftwo-wheeler collisions, in this paper,2094sample data of Changsha automobile andtwo-wheelers collision from2001to2010have been selected and arranged, and fourmain factors,including people, vehicles, roads and environmental,which contributeto the automobile and two-wheeler collision accident have been studied, categoriedand attached to accident situations. Subsequently, based on coupling system of people-vehicle-road-environment, this paper carried out comprehensive, in-depth studyof the multivariate analysis, established Bayesian network model of cars andtwo-wheelers collision, derived wheelers collision causation chain of the accident.Finally, according to the deducing results of the model,the critical control points ofthe accident was identified, the corresponding safety measures had been brought up toprevent and reduce collisions Changsha wheelers rider injury risk.The results of this study have certain guiding significance for the prevention ofChangsha automobile and two-wheelers collisions, and provided a reasonable basisfor the relevant departments to take preventive measures and scientific rationaltreatment strategies. Meanwhile, the use of Bayesian network approach has a certainsignificance for the study of collisions risk of other types.