Demonstrating the feasibility of using Wi-Fi sensors for dynamic bus-stop queue length estimation

Document Type

Journal Article

Publication Date

2024

Subject Area

place - asia, mode - bus, infrastructure - stop, ridership - demand, technology - management information systems

Keywords

Waiting time estimation, Bus stop, Passenger queue, Time-dependent arrival rates, Wi-Fi signal data

Abstract

We demonstrate that a Wi-Fi sensor located at a bus stop can be used to estimate the current queue length of passengers waiting for the bus. These crowdsourced data are hence a cost-effective source to obtain information on transit demand. Based on the observed queue lengths we further estimate the resulting waiting time for passengers for each line that serves the stop. This information can be used to evaluate the impact of operational strategies. Our experiments are conducted at a busy bus stop in Kyoto, Japan, during the peak tourist season. The estimated regression and machine learning approaches show that both randomised and non-randomised MAC addresses as well as the received signal strengths are important indicators for the queue length. We discuss the limitations of the approach if applied in other contexts with less organised crowding and more noise than in our experiment.

Rights

Permission to publish the abstract has been given by SpringerLink, copyright remains with them.

Share

COinS