engleski

Traffic Emissions Clustering Using OBD-II Dataset Based on Machine Learning Algorithms

Traffic emissions are one of the main causes of air pollution in developed urban areas. Estimating emissions patterns presents an ongoing challenge for the research and decision-making communities to detect and propose solutions for system stakeholders contributing to the pollution the most. This paper proposes a data-driven methodology for estimating the emissions clusters based on vehicle parameters that correlates to emissions. Research is based on a publicly available dataset with two main steps that include cluster number identification and the method for estimating the best-performing clustering algorithm for a given dataset. The research resulted in five emission classes based on the extracted features related to vehicle parameters and fuel consumption. The emissions clusters can be used for estimating the environmental impact of different vehicles in urban areas and producing air pollution mappings based on vehicle types.

emissions prediction ; machine learning ; emissions clustering ; OBD data ; road traffic

nije evidentirano