Short-Time Driving Style Classification and Recognition Method on Expressway

This article proposes a short-term driving style classification method that considers spatiotemporal features. By deeply mining ETC (Electronic Toll Collection) transaction data from vehicles equipped with onboard units (OBU) on the highways in Fujian Province, including historical segment speeds, segment flow rates, passage time points, vehicle types, and other data, short-term driving style features are constructed. Subsequently, the silhouette coefficient method is employed to determine the optimal number of clusters, and the K-means algorithm is used to cluster the spatiotemporal features of vehicles' driving behavior. Finally, the support vector machine is utilized to recognize driving styles. This approach accurately captures real-time features of vehicles, thereby enhancing the accuracy and reliability of driving style classification. By employing this method, more personalized driving style recognition and analysis can be provided to drivers, potentially playing a positive role in the field of intelligent driving and traffic management.