Abstract:

This paper analyzes the operating speed characteristics at horizontal curve combined with longitudinal slope and proposes prediction models for expressway, whose design speed is 100km/h. The observed speeds and the time headways are obtained, and the free flow condition is determined based on the relationship between them. Under the free flow condition, the operating speed is obtained and the relationship between operating speed and geometric alignments as radius of horizontal curve and longitudinal slope are analyzed. Lastly, the regression model, the BP neural network model and the fuzzy logic model are proposed to predict the operating speed, and the prediction accuracy is analyzed. The study results show that: When the time headway is greater than 6s, the traffic condition can be taken as free condition, which could be used to calculate operating speed; The operating speed of passenger cars are influenced significantly by radius of horizontal curve, whereas the operating speed of trucks are influenced by the gradient of upgrade; The proposed three operating speed prediction models have comparable predict capability.

Key words: traffic engineering, operating speed, free driving condition, expressway, regression analysis, neural network, fuzzy logic

摘要：

针对设计速度100km/h高速公路上平曲线与纵坡组合路段的运行速度特征及 其预测问题开展研究工作.首先，通过分析车头时距与实测车速的关系，提出了基于车头 时距的自由行驶状态判别准则.然后，基于该准则提取了自由行驶状态下的实测车速数 据，计算运行速度并分析了运行速度与平曲线半径、纵坡的关系.最后，分别建立了基于回 归分析的、基于BP神经网络的及基于模糊逻辑的运行速度预测模型，并进行了模型精度 分析.研究结果表明：(1)当车头时距大于6s时，交通处于自由行驶状态，此时的实测车速 可用于运行速度计算；(2)小客车运行速度受平曲线半径影响较大，而大货车运行速度则 随上坡坡度的增加而显著降低；(3)三个预测模型均可用于预测平曲线与纵坡组合路段的 平均运行速度.

关键词: 交通工程, 运行速度, 自由行驶状态, 高速公路, 回归分析, 神经网络, 模糊逻辑