Toward more reliable public transportation Wi-Fi origin–destination matrices: Modeling errors using synthetic noise and optical counts

Authors

• Léa Fabre
• Caroline Bayart
• Alexandre Nicolas
• Patrick Bonnel

Document Type

Journal Article

Publication Date

2025

Subject Area

place - europe, place - urban, mode - bus, ridership - behaviour, technology - passenger information

Keywords

Origin–destination matrices, Wi-Fi/bluetooth sensors, Estimation accuracy, Error modeling

Abstract

To continuously monitor mobility flows aboard public transportation, low-cost data collection methods based on the passive detection of Wi-Fi signals are promising technological solutions, but they yield uncertain results. We assess the accuracy of these results in light of a three-month experimentation conducted aboard buses equipped with Wi-Fi sensors in a sizable French conurbation. We put forward a method to quantify the error between the stop-to-stop origin–destination (O–D) matrix estimated based on Wi-Fi data and the ground truth, when the (estimated and real) volumes per boarding and alighting are known. To do so, the error in the estimated matrix is modeled by random noise. Neither additive, nor multiplicative noise replicate the experimental relation between the errors in the foregoing volumes and in the O–D matrix. Noise models that concentrate on the short O–D trips and/or the central stops better reflect the structure of the error. But only by introducing distinct uncertainties between the boarding stop and the alighting stop can we recover the asymmetry between the alighting and boarding errors, as well as the correct ratios between these aggregate errors and the O–D error. Thus, our findings give insight into the main sources of error in the Wi-Fi based reconstruction of O–D matrices. They also provide analysts with an automatic and reproducible way to control the quality of O–D matrices produced by Wi-Fi data, using (readily available) count data.

Rights

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

Comments

Transportation Research Part C Home Page:

http://www.sciencedirect.com/science/journal/0968090X

Recommended Citation

Fabre, L., Bayart, C., Nicolas, A., & Bonnel, P. (2025). Toward more reliable public transportation Wi-Fi origin–destination matrices: Modeling errors using synthetic noise and optical counts. Transportation Research Part C: Emerging Technologies, 178, 105202.