Department of Civil Engineering, Semnan University, Semnan, Iran.
Department of Civil Engineering, Science and Research Branch, Islamic Azad University, Tehran, Iran.
Accid Anal Prev. 2018 Nov;120:295-303. doi: 10.1016/j.aap.2018.08.019. Epub 2018 Sep 5.
Long-range transportation plans often involve proposals for improvements/ changes in different modes of travel. This means that modal share of trips generated at each traffic analysis zone (TAZ) by mode of travel needs to be predicted/ forecasted for safety evaluation purposes. The objective of this research study is to develop a series of aggregate crash prediction models (ACPMs) that relate with the modal split step of the conventional four-step demand models.
The models are developed utilizing network and vehicular, socio-economical, trip production/attraction and trip frequencies by mode at TAZ-level as explanatory variables in a generalized linear regression with the assumption of a negative binomial error structure. Crash frequencies are split into total crashes (TC) and severe crashes (SC).
The models prove promising in estimating crash frequencies upon changes in modal shares, which is essential in safety assessment of alternate transportation demand management (TDM) scenarios. Trips made in car, bus, and bus Service mode became significant in the estimated TC and trips made in car, taxi, school service, bus service and moped mode became significant in the estimated SC ACPMs.
The ACPMs may be used from two different points of view. First and most appropriate use is to consider these as tools to forecast future crash frequencies and develop long-term plans to counteract. In the second point of view, ACPMs act as the primary planning tool to identify how any increase in a specific mode-ridership will contribute to crash frequencies. This is of great interest in developing plans that involve increased use of a specific mode.
As modal shares are forecasted in certain years into the future by the modal split step of demand modeling, crash frequencies could also be forecasted and safety implications of mobility improvement scenarios (e.g. increased number of trips by bus, car, etc.) would be evaluated.
长途交通规划通常涉及不同交通方式的改进/变更提案。这意味着需要预测/预估每个交通分析区（TAZ）的每种交通方式产生的出行方式的模态分担率，以便进行安全评估。本研究的目的是开发一系列与传统四步骤需求模型的模态划分步骤相关的综合碰撞预测模型（ACPM）。
模型的开发利用了网络和车辆、社会经济、出行生成/吸引和出行频率，在 TAZ 级别按模式作为解释变量，在广义线性回归中假设负二项式误差结构。碰撞频率分为总碰撞（TC）和严重碰撞（SC）。
模型在估计模式份额变化时的碰撞频率方面表现出良好的效果，这对于替代交通需求管理（TDM）情景的安全评估至关重要。在估计 TC 时，汽车、公共汽车和公共汽车服务模式的出行变得显著，在估计 SC 时，汽车、出租车、学校服务、公共汽车服务和轻便摩托车模式的出行变得显著。
ACPM 可以从两个不同的角度来看待。首先也是最合适的用途是将这些视为预测未来碰撞频率并制定长期计划以应对的工具。从第二个角度来看，ACPM 作为主要规划工具，可以确定特定模式的出行量增加将如何导致碰撞频率增加。这对于制定涉及特定模式使用增加的计划非常重要。
由于需求建模的模态划分步骤可以预测未来某些年份的模态分担率，因此也可以预测碰撞频率，并评估改善出行方案（例如增加公共汽车、汽车等的出行次数）的安全影响。
