ShibuyaSocial: Multi-scale Model of Pedestrian Flows in Scramble Crossing

TL;DR

This paper introduces a learning-based multi-scale pedestrian flow model at Shibuya crossing, combining global route choice and local collision avoidance, validated through video data and simulations.

Contribution

It presents a novel integrated model that combines global route selection with local collision avoidance using an Attention mechanism, addressing limitations of prior local-only models.

Findings

Model accurately predicts pedestrian behaviors at Shibuya crossing.

Simulation results align with real-world pedestrian flow patterns.

The approach enhances safety and environment design in urban spaces.

Abstract

This paper presents a learning-based model of pedestrian flows that integrates multi scale behaviors such as global route selection and local collision avoidance in urban spaces, particularly focusing on pedestrian movements at Shibuya scramble crossing. Since too much congestion of pedestrian flows can cause serious accidents, mathematically modeling and predicting pedestrian behaviors is important for preventing such accidents and providing a safe and comfortable environment. Although numerous studies have investigated learning-based modeling methods, most of them focus only on the local behavior of pedestrians, such as collision avoidance with neighbors and environmental objects. In an actual environment, pedestrian behavior involves more complicated decision making including global route selection. Moreover, a state transition from stopping to walking at a traffic light should be considered simultaneously. In this study, the proposed model integrates local behaviors with global route selection, using an Attention mechanism to ensure consistent global and local behavior predictions. We recorded video data of pedestrians at Shibuya scramble crossing and trained the proposed model using pedestrian walking trajectory data obtained from the video. Simulations of pedestrian behaviors based on the trained model qualitatively and quantitatively validated that the proposed model can appropriately predict pedestrian behaviors.