Modelling Road Congestion Using a Fuzzy System and Real-World Data for Connected and Autonomous Vehicles

Road congestion is estimated to cost the United Kingdom £307 billion by 2030. Furthermore, congestion contributes enormously to damaging the environment and people's health. In an attempt to combat the damage congestion is causing, new technologies are being developed, such as intelligent infrastructures and smart vehicles. The aim of this study is to develop a fuzzy system that can classify congestion using a real-world dataset referred to as Manchester Urban Congestion Dataset, which contains data similar to that collected by connected and autonomous vehicles. A set of fuzzy membership functions and rules were developed using a road congestion ontology and in conjunction with domain experts. Experiments are conducted to evaluate the fuzzy system in terms of its precision and recall in classifying congestion. Comparisons are made in terms of performance with traditional classification algorithms decision trees and Naïve Bayes using the Red, Amber, and Green classification methods currently implemented by Transport for Greater Manchester to label the dataset. The results have shown the fuzzy system has the ability to predict road congestion using volume and journey time, outperforming both decision trees and Naïve Bayes.