Identifying intracity trip ends of hazardous goods trucks from GPS trajectories

With the global GPS positioning system's advancement, the acquisition of extensive GPS data from hazardous goods trucks has become efficient. Identifying endpoints of freight trips from hazardous goods truck GPS data stands as a crucial research problem. While prior studies have proposed various methods for identifying trip endpoints from different perspectives, these methods have predominantly relied on singular time thresholds, speed thresholds, or spatial distance thresholds, thus leading to an incomplete consideration of the complex travel characteristics of hazardous goods trucks. We propose a clustering algorithm based on speed and spatiotemporal factors to address this limitation. This method comprehensively considers speed, time, and spatial aspects. Initially, we apply a speed threshold to obtain initial clusters of stop points. Subsequently, neighboring stop clusters are merged based on the maximum distance threshold (M_dist) and the maximum time threshold (M_dura) to filter out short-duration pseudo-stop points that do not meet the time threshold. Additionally, we leverage Point of Interest (POI) data relevant to freight transportation to aid in eliminating long-duration pseudo-stop points. Finally, we validate the identified endpoints using order data and satellite maps, achieving an accuracy rate of 88.12%. Our approach extracts hidden information and travel characteristics from hazardous goods truck GPS data, enabling the visualization of spatial distributions of urban hazardous goods truck freight trips. This method provides data support for the sustainable development of the hazardous goods transportation industry and planning transportation freight networks, offering significant practical application value.