Dynamic Assessment of Pedestrian Level of Service to Support Spatial Retrofit and Operational Interventions in an Underground Transport Hub: A Case Study of Chongqing Shapingba Station

Underground transport hubs must accommodate concentrated passenger arrivals in constrained multilevel spaces. Aggregate demand and period-average density offer limited guidance on whether crowding calls for geometric modification, facility reconfiguration, or circulation management. This study develops a field-data-informed framework combining agent-based simulation with dynamic level-of-service (LOS) assessment based on Fruin’s criteria to support spatial diagnosis and intervention selection. A MassMotion model of Chongqing Shapingba Station was developed using field counts, observed circulation patterns, and train-operation data. Experienced spatial density, maximum spatial density, LOS D–F duration, and individual social cost were used to identify hotspots and compare three interventions. Although pedestrian counts were higher during 08:00–09:00, crowding was more pronounced during 12:00–13:00, when concentrated high-speed rail arrivals generated LOS D–F conditions in the high-speed rail exit hall and around the Metro Line 9 security-check and fare-gate areas. Corridor widening produced the largest reductions in the extent and persistence of high-density conditions. Security-exempt transfer reduced total individual social cost by 8.19%, whereas fare-gate direction reversal produced a smaller improvement. By distinguishing geometric constraints, processing bottlenecks, and recurring flow conflicts, the framework supports targeted spatial retrofit and operational interventions in existing underground transport hubs.