Driver Behavior Profiling: Clustering Techniques for Enhanced Road Safety and Personalized Mobility

This paper explores the use of connected vehicle datasets to detect and classify driver behaviors, aiming to identify distinct driving profiles through advanced clustering techniques. By analyzing real-world driving data from vehicle sensors and GPS systems, we uncover patterns that characterize different types of drivers. Using machine learning algorithms such as k-means, DBSCAN, and Agglomerative clustering, we identify several driving profiles based on behavioral traits like aggressiveness, caution, and driving habits in various road conditions. Our findings offer valuable insights for improving road safety, developing personalized driver assistance systems, and enhancing traffic management strategies. Additionally, these insights can be applied to create more accurate insurance models, where customers pay based on how they drive rather than traditional demographic factors. This "pay-as-you-drive" model offers a more individualized approach to risk assessment and can incentivize safer driving practices. This research highlights the potential of connected vehicle data to transform the future of transportation by enabling personalized and adaptive driving experiences.