Reproduce “Realistic Urban Traffic Generator using Decentralized Federated Learning for the SUMO simulator” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Alberto Bazán-Guillén, Carlos Beis-Penedo, Diego Cajaraville-Aboy, Pablo Barbecho-Bautista, Rebeca P. Díaz-Redondo, Luis J. de la Cruz Llopis, Ana Fernández-Vilas, Mónica Aguilar Igartua and 1 more
DesRUTGe, a system developed at the Universitat Autònoma de Barcelona, integrates Deep Reinforcement Learning with Decentralized Federated Learning to generate realistic 24-hour urban traffic patterns for the SUMO simulator. The framework was validated using real-world data from Barcelona's traffic network, achieving a nearly 3x improvement in accuracy over standard SUMO tools while enhancing privacy by eliminating centralized data collection.
Key insights
- Decentralized Federated Learning enables collaborative urban traffic modeling without compromising data privacy, as raw mobility data never leaves local nodes.
- Spatial proximity (Geographic Neighbors strategy) is more effective than traffic pattern similarity for model sharing in urban environments, leading to faster convergence and lower errors.
- Deep Reinforcement Learning, specifically Proximal Policy Optimization (PPO), can effectively learn to dynamically adjust vehicle injection rates to match complex, time-varying traffic intensities.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.