Modelling Traffic Conditions in Developing Urban Areas: A Combined Approach of Explainable Artificial Intelligence and Mobile Crowd-Sourcing

Traffic congestion remains a persistent and challenging issue in major urban centers, particularly in developing countries where transportation infrastructure struggles to keep pace with rapidly growing demand. This paper proposes innovative solutions to address this issue by introducing mobile crowd-sourcing based approaches for traffic condition estimation. A scalable framework is outlined for the efficient collection, integration, and analysis of traffic-related data contributed by mobile crowds. Furthermore, the paper addresses critical issues related to reliably predicting traffic conditions where prediction interpretation is unavailable, utilizing an machine learning model backed by Explainable Artificial Intelligence. By leveraging the explainability of the model, valuable insights can be extracted to enhance traffic conditions and improve overall social welfare.