Quantifying Carbon Emissions: Integrating Machine Learning with IoT for Traffic-Aware Emission Scores

The current nature of society and its dependency on fossil fuels contribute to a consistent trend in the increase of global CO<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</inf> emissions. Since 1990 to the present, global CO‚ emissions have increased by more than 60 percent. The amount of CO‚ emitted by a average passenger car is typically around 4.6 metric tonnes per year. This paper aims to predict traffic conditions categorised by the following congestion levels: low, normal, and high congestion levels, and subsequently calculate the carbon emission score. This is accomplished through the integration of two critical components of computer science: the Internet of Things (IOT) and machine learning (ML). The study also aligns with the SDG 7 (Affordable and Clean Energy) and the SDG 11 (Sustainable Cities and Communities) by seeking to create efficient and eco-friendly traffic systems by establishing a standard way of comparing carbon emissions by looking at the external environment and the vehicle data during a trip. This is achieved by analysing traffic data and total carbon emissions to suggest a carbon emission score for all cars, allowing for a more precise comparison. Several machine learning models, like linear regression, random forest classifier, XGBoost, and AdaBoost, were used to predict traffic congestion. The findings indicate that XGBoost surpasses other models with accuracy of 98.9%. The project’s low cost and minimal resources is to facilitate user adoption and implementation. This study aims to promote strategies that contribute to environmental sustainability and advance efforts toward a cleaner, more sustainable future.