Dynamic management tool for improving passenger experience at transport interchanges

Resumen

This study proposes a methodology that integrates real-time data and predictive modeling to identify the passenger flow and occupancy levels within a multimodal transport hub. This tool enables the implementation of control and planning strategies to ensure a high Level of Service (LOS). The tool is based on a Long Short-Term Memory (LSTM) model and heterogeneous data sources, including an Automatic Passenger Counting (APC) system, which are utilized to estimate the real-time passenger flow and area occupancy. The Module A of the Moncloa Interchange in Madrid is the case study, and the results reveal that transport-dedicated zones have higher occupancy levels. Methodologically, time series data were standardized to a uniform frequency to ensure consistency, and the training set consisted of seven months of available data. The model performs better in high-occupancy zones. Despite maintaining a LOS A, some periods experience temporary congestion. These findings indicate that the variations in occupancy levels influence the service quality and highlight the essential role of dynamic interchange management. Tailored operational strategies can optimize the service levels and improve the user experience by anticipating congestion through predictive modeling. This can help enhance public transport's attractiveness, minimize the perceived transfer penalties, make transfers more efficient, and reinforce transport hubs' role in sustainable urban mobility.