Validation of agent-based passenger movement modeling for railway stations subject to social distancing during the COVID-19 pandemic

The COVID-19 pandemic has caused unprecedented difficulties keeping passenger transport running while social distancing has constrained pedestrian movement. To support railway operation an agent based simulation validated against UK mainline railway station CCTV footage has been developed. The simulation quantifies train alighting and boarding flow times for different degrees of distancing, rolling stock types, platform size, and passenger numbers. Real-world social distancing behaviour was represented including compromises people make on distancing to achieve their goals. Flow times modelled and those measured from CCTV were within 10%. Relative to a baseline without social distancing, 1 m distancing was predicted to make only marginal difference to passenger flow times, primarily because of passenger behaviour to compromise on distancing at the rate determining door location. For 2 m social distancing significantly extended passenger alighting and boarding flow times were predicted for busier services (2.2-2.8 times baseline depending on rolling stock type). These increases in flow time are only predicted to begin when the combined total of boarding and alighting passengers exceeds 10-15 per door. The model has application on transport systems worldwide in avoiding unmaintainable timetables, and in reducing incentives for social distancing compromise when distancing forms a component of suppressing virus transmission.