关键词: 交通工程, 交通安全, 摩托车事故, 巢式logit模型, 单车事故, 事故严重程度

Abstract: 1526 motorcycle crashes in Indiana in America from 2013 to 2015 were used to study the factors influencing the severity of motorcycle crash accidents on roadway segments. Considering 21 independent variables from 5 aspects such as drivers characteristics, vehicle attributes, road conditions, environment and accident characteristics, a nested logit model was constructed to account for the correlation between severity categories (fatality, incapacitating injury, non-incapacitating injury, property damage only). The model parameters were estimated by Full Information Maximum Likelihood (FIML). The estimation results show that the overall fitness of the proposed model is good. The results show that gender of motorcyclist, age of motorcyclist, helmet usage, alcohol use, ejection or not, speeding or not, running off the road or not, passenger on motorcycle, age of vehicle, wet pavement or not, horizontal curve, speed limit, months, light conditions, rural areas or not, motorcycle collisions (pole, curb, culvert, guardrail, tree) are associated with the crash injury severity at 0.1 significance level. Factors associated with a substantially higher probability of fatality include ejection, speeding, motorcycle age over 10, running off the road, accidents occurred in April, July, night, accidents occurred in rural areas and motorcycle collisions with poles. These findings can help transportation agencies develop appropriate strategies to reduce injury severity resulting from single-motorcycle accidents on roadway segments, and can also provide reference for further analysis of motorcycle accidents in China in the later period.

Key words: traffic engineering, traffic safety, motorcycle crash, nested logit model, single vehicle crash, crash injury severity