Pedestrian Flow: Monitoring and Prediction

The worldwide lockdowns during the first wave of the COVID-19 pandemic had an immense effect on the public space. The events brought up an opportunity to redesign mobility plans, streets, and sidewalks, making cities more resilient and adaptable. This paper builds on previous research of the authors that focused on the development of a graphene-based sensing material system applied to a smart pavement and utilized to obtain pedestrian spatiotemporal data. The necessary steps for gradual integration of the material system within the urban fabric are introduced as milestones toward predictive modeling and dynamic mobility reconfiguration. Based on the capacity of the smart pavement, the current research presents how data acquired through an agent-based pedestrian simulation is used to gain insight into mobility patterns. A range of maps representing pedestrian density, flow, and distancing are generated to visualize the simulated behavioral patterns. The methodology is used to identify areas with high density and, thus, high risk of transmitting airborne diseases. The insights gained are used to identify streets where additional space for pedestrians is needed to allow safe use of the public space. It is proposed that this is done by creating a dynamic mobility plan where temporal pedestrianization takes place at certain times of the day with minimal disruption of road traffic. Although this paper focuses mainly on the agent-based pedestrian simulation, the method can be used with real-time data acquired by the sensing material system for informed decision-making following otherwise-unpredictable pedestrian behavior. Finally, the simulated data is used within a predictive modeling framework to identify further steps for each agent; this is used as a proof-of-concept through which more insights can be gained with additional exploration.