Abstract

In the domain of smart mobility, Artificial Intelligence (AI) methodologies play a pivotal role, offering substantial potential for transformative impact. This project focuses on the development of a real-time ecological map of road traffic, designed to enable electric vehicles (EVs) and thermal vehicles (TVs) to visualize energy consumption costs and CO2 emissions across various road segments. In urban settings, vehicle-related emissions significantly contribute to environmental degradation, underscoring the need for a detailed analysis of vehicle operational dynamics across diverse road infrastructures. This study introduces the Vehicle Activity Dataset (VAD), a novel and comprehensive dataset aimed at evaluating vehicle emissions and fuel consumption in relation to their operational contexts. Compiled from extensive real-world driving scenarios, VAD integrates emission data gathered via an industrial Portable Emission Measurement System (PEMS), road imagery captured by an RGB camera, and the identification of various object classes within these images. The primary goal of VAD is to elucidate the interplay between vehicle emissions and the multitude of objects encountered on roadways. Empirical evaluations conducted in authentic road traffic conditions across multiple studies affirm the robustness and reliability of the dataset.