Artificial Neural Networks for Forecasting Passenger Flows on Metro Lines

Area of Science:

• Transportation Engineering
• Artificial Intelligence
• Operations Research

Background:

• Intelligent Transport Systems (ITSs) rely on accurate user flow data for effective management.
• Forecasting passenger flows on network links is crucial for ITS strategy implementation.
• Real-time data from sensors can significantly enhance transportation network predictions.

Purpose of the Study:

• To propose and evaluate Artificial Neural Networks (ANNs) for forecasting metro onboard passenger flows.
• To estimate passenger counts on track sections using historical turnstile data.
• To provide a data-driven approach for improving metro operational efficiency.

Main Methods:

• Utilizing Artificial Neural Networks (ANNs) to model passenger flow dynamics.
• Training ANNs with simulation data generated from a dynamic rail line loading procedure.
• Employing metro station turnstile data, aggregated over short periods, as input for the models.

Main Results:

• The proposed ANN approach demonstrated satisfactory precision in forecasting passenger flows on metro sections.
• The method successfully estimated onboard passengers based on turnstile entry counts.
• The approach was validated on a real-world case study: Line 1 of the Naples metro system.

Conclusions:

• ANNs offer a viable and precise method for forecasting metro passenger flows.
• This technique can enhance the real-time management capabilities of Intelligent Transport Systems.
• The study provides a practical application of AI in public transportation network optimization.