Public Traffic Congestion Estimation Using an Artificial Neural Network

Abstract

Alleviating public traffic congestion is an efficient and effective way to improve the travel time reliability and quality of public transport services. The existing public network optimization models usually ignored the essential impact of public traffic congestion on the performance of public transport service. To address this problem, this study proposes a data-based methodology to estimate the traffic congestion of road segments between bus stops (RSBs). The proposed methodology involves two steps: (1) Extracting three traffic indicators of the RSBs from smart card data and bus trajectory data; (2) The self-organizing map (SOM) is used to cluster and effectively recognize traffic patterns embedded in the RSBs. Furthermore, a congestion index for ranking the SOM clusters is developed to determine the congested RSBs. A case study using real-world datasets from a public transport system validates the proposed methodology. Based on the congested RSBs, an exploratory example of public transport network optimization is discussed and evaluated using a genetic algorithm. The clustering results showed that the SOM could suitably reflect the traffic characteristics and estimate traffic congestion of the RSBs. The results obtained in this study are expected to demonstrate the usefulness of the proposed methodology in sustainable public transport improvements.