EURO 2024 Copenhagen
Abstract Submission

EURO-Online login

1858. High-resolution Platoon Prediction for Coordinated Traffic Control along Urban Arterials

Invited abstract in session TD-56: Advancing mobility towards sustainable solutions II, stream Transportation.

Tuesday, 14:30-16:00
Room: S04 (building: 101)

Authors (first author is the speaker)

1. Zakir Farahmand
Civil Engineering and Management, University of Twente
2. Oskar Eikenbroek
Transport Engineering and Management, University of Twente
3. Eric van Berkum
Civil Engineering, University of Twente

Abstract

Smooth progression of platoons of vehicles along urban arterials can significantly contribute to environmental objectives, especially when heavy-duty vehicles are involved. Generally, signal controls such as fixed-time, vehicle-actuated and adaptive aim to minimise delays, which may conflict with reducing the number of stops to improve air quality and GHG emissions. Overcoming this challenge requires dynamically synchronising the signal timings of a series of intersections with approaching vehicles to simultaneously reduce delays and the number of stops. A fundamental prerequisite of such controls is real-time prediction of platoon dynamics including forming, dispersion, stopping and arrival after departing from upstream. Existing coordination methods usually focus on flow estimation at aggregate levels, losing critical information for multi-criteria signal operations. Our study introduces a novel method for predicting platoons’ dynamics with high resolution under mixed-traffic environments. Utilizing various data sources, including loop detectors, vehicle-to-infrastructure (V2I) communication and floating car data, we employ attention-based graph neural networks to predict platoon arrival and stop times while considering current and anticipated signal phases. Experiments conducted on real-world datasets from an arterial in Enschede, the Netherlands, reveal the suitability of our approach for coordinated signal controls to reduce the number of stops and enable green waves.

Keywords

Status: accepted

Back to the list of papers