ABSTRACT

Vehicle emission analysis currently faces a trade-off between easy-to-use, low-accuracy macroscopic models, and computationally intensive, high-accuracy microscopic models. In this study, we develop a surrogate model that leverages microscopic traffic and emission simulations to predict link-level emission rates. The input variables are obtained by aggregating 1 Hz simulated vehicle trajectories into hourly traffic condition factors (e.g., link average/variation of speed, truck fleet percentage, road grade, etc.). The emission ground truth data are generated using the Motor Vehicle Emission Simulator (MOVES) opmode-based analysis module. We explore different parameter and machine learning model structures to establish the statistical relationship of the input variables and the link-level emission rates. We demonstrate the ability of our model to accurately estimate vehicle-related emissions by using the Columbia, South Carolina road network as an example. This model can serve as a high-level planning tool to assess the impacts of emissions from transportation projects.

Implications: Vehicle emission analysis is facing trade-offs between easy-to-use macroscopic emission models with low accuracy and computationally intensive microscopic models with high accuracy. Existing studies attempted to cope with the trade-off by pre-selecting representative emission rates but are still subject to the risk of not considering differentiated traffic patterns by using single emission rate. To fill in the knowledge gap in the literature, we develop a surrogate approach that fully integrates driving trajectories of heterogenous traffic patterns into a link-level emissions estimation model considering road characteristics. The model can achieve high accuracy and utilize publicly available traffic data in vehicle emission prediction. We apply the proposed model in a middle size city road network and demonstrate its capability to capture and quantify the impacts of traffic patterns on link-level vehicle-related emissions. Additionally, the proposed model can serve as a sketch planning tool for researchers and transportation air quality practitioners to quickly assess bounds of emissions benefits due to traffic operational and transportation strategies.

摘要

车辆排放分析目前面临着易于使用的低精度宏观模型和计算密集型高精度微观模型之间的权衡。在这项研究中，我们开发了一个替代模型，该模型利用微观交通和排放模拟来预测链路级排放率。输入变量是通过将 1 Hz 模拟车辆轨迹聚合为每小时交通状况因素（例如，链路平均/速度变化、卡车车队百分比、道路坡度等）而获得的。排放地面实况数据是使用基于机动车辆排放模拟器 (MOVES) 操作模式的分析模块生成的。我们探索不同的参数和机器学习模型结构，以建立输入变量和链路级排放率的统计关系。我们以南卡罗来纳州哥伦比亚市的道路网络为例，证明了我们的模型能够准确估计与车辆相关的排放。该模型可用作评估交通项目排放影响的高级规划工具。

影响：车辆排放分析正面临着易于使用的低精度宏观排放模型和高精确度的计算密集型微观模型之间的权衡。现有的研究试图通过预先选择有代表性的排放率来应对这种权衡，但仍然面临着不通过使用单一排放率来考虑差异化交通模式的风险。为了填补文献中的知识空白，我们开发了一种替代方法，将异构交通模式的驾驶轨迹完全整合到考虑道路特征的链路级排放估计模型中。该模型可以实现高精度并在车辆排放预测中利用公开可用的交通数据。我们将所提出的模型应用于中等规模的城市道路网络，并展示了其捕捉和量化交通模式对链路级车辆相关排放的影响的能力。此外，所提出的模型可以作为研究人员和交通空气质量从业人员的草图规划工具，以快速评估由于交通运营和交通战略而产生的排放效益范围。