Real-time, on-board crowding estimation in public transport networks with multiple lines using non-exhaustive Automatic Passenger Counting data

Activity: Talk or presentation typesConference presentation

Description

Accurate information about passenger volumes and flows in public transport is important for the efficient operation, management, and evaluation of the network. Passengers’ comfort of travel is a major criterion for choosing public transport against less sustainable modes and the prevention of crowding inside vehicles is a challenging task for managers and operators of public transport services. The avoidance of crowds became even more critical during COVID-19, which highlighted the need for preparedness in terms of a proper provision of information on crowding phenomena. In recent years, information about passenger volume on-board public transport vehicles is commonly derived from Automatic Passenger Count data. Such data are often incomplete and there is a critical need for methods to estimate the missing records. An existing study developed a Kalman filter-based scheme for estimating the number of passengers on-board public transport vehicles, employing non-exhaustive real-time Automatic Passenger Counting data. The current study builds upon this study and extends it in order to allow estimations for networks with multiple common lines per station. The accuracy and reliability of the estimation are evaluated through application to the commuter train network of Helsinki, Finland, and the results suggest that the proposed method is able to deliver good estimation accuracy in terms of the number of passengers boarding, alighting, and, ultimately, comfort Levels of Service.
Period2024
Event titleTransportation Research Board Annual Meeting
Event typeConference
Conference number103
LocationWashington, United States, District of ColumbiaShow on map
Degree of RecognitionInternational

Keywords

  • public transport; on-board comfort; Automatic Passenger Counting; Kalman filter