Identifying intercity freight trip ends of heavy trucks from GPS data

• Identify truck stops by capturing GPS trajectory characteristics. • Determine time thresholds by capturing the temporal characteristics of truck activities. • Detect freight-related POI boundaries based on mean-shift algorithm. • The accuracy of our proposed method is significantly improved compared to the benchmark methods. • Analyze the spatiotemporal distribution of intercity freight trips of heavy trucks in China. The intercity freight trips of heavy trucks are basic data for transportation system planning and management. In recent decades, extracting intercity freight trips from GPS data has gradually become the main alternative to traditional surveys. Identifying freight trip ends (origin and destination) is the first task in trip extraction. Although many trip end identification methods have been proposed in previous studies, most of these studies subjectively determined key parameters and ignored the complex characteristics of truck trajectory and freight activities. In this paper, we propose a data-driven trip end identification method based on massive GPS data of heavy trucks in China. First, we capture heavy truck trajectory characteristics under the influence of GPS drift to identify truck stops from GPS data. Second, we analyze the temporal characteristics of truck activities and use freight-related point-of-interest (POI) data and highway network GIS data to identify valid trip ends from truck stops. The results of method validation suggest that the accuracy of our proposed method is significantly improved in comparison with the benchmark methods. We further extract intercity freight trips from the identified trip ends and analyze the spatiotemporal characteristics of intercity freight trips in China.