Estimating bus passenger volume based on a Wi-Fi scanner survey. (July 2020)

• Record Type: Journal Article
• Authors: Hidayat, Arief; Terabe, Shintaro; Yaginuma, Hideki
• Is Part Of: Transportation research interdisciplinary perspectives. Volume 6(2020)
• Publication Date: 2020-07
• Subjects: Wi-Fi scanner -- Passenger volume -- Bus passenger -- Estimated passenger; Transportation -- Periodicals; 388.05
• DOI: 10.1016/j.trip.2020.100142
• Languages: English
• ISSNs: 2590-1982

Abstract

Abstract: The main objective of this study is to estimate bus passenger volume based on a Wi-Fi scanner transportation survey. The study proposes a new data cleaning procedure to characterize bus passenger volume using a combination of media access control (MAC) address and global positioning system (GPS) data. We used a Wi-Fi scanner to detect the MAC addresses of individual bus passengers. The Wi-Fi scanner, used as a tool to capture passenger device data, can engage in a probe request mode to capture the MAC addresses of mobile devices or other Wi-Fi-enabled modalities without connecting to the internet. The approach proposed in this study can yield the passenger volume outputs for various bus route sections. A comparison of the passenger volume results obtained from Wi-Fi data and ground truth data indicates that the number of passengers determined from the former is less than that from the latter. The correlation between Wi-Fi estimation and ground truth is 0.78, and the trend line in both methods is similar. Therefore, the cleaning procedure proposed in this study can effectively clean raw Wi-Fi data to extract passenger volume data. Highlights: A new procedure applied to clean the non-passenger data from the total Wi-Fi access data. The estimated speed of the non-passengers' whose Wi-Fi access was cleaned, was calculated based on route, time and GPS data. The estimated bus passenger volume from the cleaned Wi-Fi access closely resembles the actual number of passengers; the correlation is 0.78.