The rapid expansion of the logistics industry has led to sustained growth in freight demand across regions. Analyzing the factors that influence truck travel demand and extracting loading/unloading locations from truck trajectory data can offer a theoretical foundation and policy support for freight logistics planning and management. Accordingly, this study is structured around two core tasks: identifying truck trip endpoints and forecasting truck travel demand, aiming to thoroughly examine urban freight travel patterns and their influencing factors. First, truck GPS data is integrated with a non-parametric Lorenz curve fitting method, road network data, and Areas of Interest (AOI) to precisely identify truck route endpoints. Subsequently, a Multi-Feature Dynamic GCN-LSTM model is constructed to predict truck travel demand by comprehensively considering multi-dimensional regional characteristics. Using Xi’an, China, as a case study, the results show that the proposed framework provides a more precise spatial distribution of truck trips and effectively forecasts truck travel demand in the city. This research has significant implications for enhancing urban freight efficiency, optimising the layout and facilities of urban transport networks, developing evidence-based transport policies, and promoting logistics development.
Cheng et al. (Fri,) studied this question.