A cellular automata approach for modelling pedestrian-vehicle mixed traffic flow in urban city

TL;DR

This paper introduces a multi-grid cellular automata model combining vehicle and pedestrian dynamics to simulate urban mixed traffic flow, highlighting the impact of pedestrian intrusion and traffic management strategies.

Contribution

It presents a novel integrated cellular automata framework with improved pedestrian and vehicle models for accurate mixed traffic simulation.

Findings

High simulation accuracy achieved

Pedestrian intrusion significantly affects traffic flow

Lower speeds and wider sidewalks reduce conflicts

Abstract

In urban streets, the intrusion of pedestrians presents significant safety challenges. Modelling mixed pedestrian-vehicle traffic is complex due to the distinct motion characteristics and spatial dimensions of pedestrians and vehicles, making unified modelling difficult, with few studies addressing these issues. This paper employs a multi-grid cellular automata model to bridge the gap between vehicle and pedestrian models. An Improved Kerner-Klenov-Wolf (IKKW) model and a pedestrian motion model that incorporates Time-To-Collision (TTC) are introduced. Both models update the spatial motions of vehicles and pedestrians uniformly. Empirical analysis indicates that the model achieves high simulation accuracy. This model effectively illustrates the impact of pedestrian intrusion within mixed traffic scenario. The fundamental diagram of heterogeneous traffic reveals substantial differences, highlighting the effects of pedestrian intrusion on traffic flow states and identifying six phase regions in mixed traffic. Additionally, this paper examines conflicts between pedestrians and vehicles under varying speed limits and sidewalk widths, demonstrating that lower speeds and broader sidewalks significantly reduce the frequency of pedestrian-vehicle conflicts. Notably, the frequency of peak conflicts at a vehicle speed limit of 60.48 km/h is more than three times higher than at 30.24 km/h. This model offers a potential approach to studying mixed traffic flows and exhibits substantial scalability.