Chen, Hongxu (Zhejiang University), Zhou, Qishen (Zhejiang University), Zhang, Bugao (ENJOYOR Technology CO.,LTD), Liu, Hongjun (Zhejiang University), Cai, Zhengyi (Zhejiang University), Zheng, Zhu (Zhejiang University), Hu, Simon (Zhejiang University)

Urban Traffic Carbon Emission Modeling Via Random Forest: Feature Importance and Effects of Sensors

Scheduled for presentation during the Regular Session "S07b-Smart Infrastructure and Data-Driven Sensing for Intelligent Mobility" (WE-EA-T7), Wednesday, November 19, 2025, 14:30−14:50, Coolangata 1

2025 IEEE 28th International Conference on Intelligent Transportation Systems (ITSC), November 18-21, 2025, Gold Coast, Australia

This information is tentative and subject to change. Compiled on October 19, 2025

Abstract

With the rapid increase in vehicle ownership, urban traffic has become a major source of carbon emissions, highlighting the urgent need for effective emission control strategies. Accurately estimating traffic-related emissions at high spatial and temporal resolution is essential for effective mitigation strategies, yet remains challenging due to data constraints. In this study, we propose a mesoscopic carbon emission model based on the comprehensive and high-resolution pNEUMA trajectory dataset, achieving a ideal estimation error of 13.8%. Using SHapley Additive exPlanations (SHAP), we interpret the model and quantify the influence of key features such as vehicle density, road segment length, average speed, average acceleration, and acceleration variance. Furthermore, the functional relationships between emission levels and average speed, average acceleration, and acceleration variance are given. We also simulate data from real-world sensor environments and demonstrate that integrating Connected Vehicle (CV) data with Automatic Number Plate Recognition (ANPR) data significantly improves estimation accuracy. Finally, we evaluate the model's generalizability.