Predicting Passenger Flow Using Graph Neural Networks with Scheduled Sampling on Bus Networks

Predicting short-term passenger flows in bus networks is crucial to improving the overall performance of such systems and increasing their attractiveness. This study develops a graph neural network-based framework for multi-step passenger flow prediction specifically designed for bus networks to capture their unique characteristics. We propose the Multi-step Multi-component Graph Convolutional Long Short-Term Memory (Multi-GCN-LSTM) model, which uses 1) a proximity matrix in addition to an adjacency matrix to consider the effects of vehicular traffic and link-level distances; 2) Scheduled Sampling for multi-step prediction, which prevents error propagation across prediction steps; and 3) a novel fusion mechanism for considering time-varying spatial and temporal correlations among passenger flow data based on recent, daily, and weekly travel patterns. This model is validated using real-world data collected from the Laval bus network. Also, benchmarking the established model against state-of-the-art baselines indicated its competency.

I. Introduction

Bus transportation systems are essential parts of urban transportation networks. They can reduce the share of private cars and thus reduce traffic congestion, air pollution, and fuel consumption. The relative crowdedness of buses in the network plays a significant role in the optimal operation and increasing the attractiveness of these systems. Knowing about passenger flows, which in this study means the number of people in the bus, on a specific bus or route at a short-term horizon can help decision-makers and transit operators to control the passenger outflows and inflows at bus stops. Moreover, informing passengers about the crowdedness of arriving buses in the network will help them decide when and what services or modes of transport to use.