Deep Reinforcement Learning-based Traffic Signal Control for Sustainable Urban Mobility

Strategic Insights -

Researchers at the Beijing University of Technology have developed a novel traffic signal control method based on Deep Reinforcement Learning (DRL) to improve traffic efficiency and reduce carbon emissions in urban transportation systems. The method integrates with Cooperative Vehicle-Infrastructure Systems (CVIS) and a doubly Day-to-Day (DTD) Dynamic Traffic Assignment model to optimize signal timings and traveler departure times and route choices. A case study on the Sioux Falls network shows that the DRL-based traffic signal strategy outperforms traditional fixed-time control strategies, achieving CO2 emission reductions of 21% to 27% in various scenarios.

Key Takeaways:

  • The proposed DRL-based method improves signal control efficiency and carbon emission reduction under complex traffic conditions, with up to 54.9% emissions reduction on high-traffic Link 28 (S-V) in the S-V scenario.
  • The method is integrated with CVIS to provide real-time traffic flow data that informs the DRL model's optimization of signal timings.
  • The DTD model adjusts traveler departure times and route choices for significant emission reductions, showcasing the potential of DRL in low-carbon traffic management.
  • The research emphasizes the need for sustainable development of future traffic systems, offering practical insights and robust solutions for emission reduction and efficient traffic management in UTS.
  • The study has been peer-reviewed and published in the journal Applied Energy.

Statistics:

  • CO2 emission reductions of 21% to 27% achieved by the DRL-based traffic signal strategy in various scenarios.
  • Up to 54.9% emissions reduction on high-traffic Link 28 (S-V) in the S-V scenario.
  • The DRL-based method significantly improves signal control efficiency and carbon emission reduction under complex traffic conditions.

Sources:

  • NewsRx. New Findings from Beijing University of Technology Describe Advances in Global Warming and Climate Change (The Impact of Deep Reinforcement Learning-based Traffic Signal Control On Emission Reduction In Urban Road Networks Empowered By ...). Global Warming Focus. July 21, 2025; p 1660.
  • The Impact of Deep Reinforcement Learning-based Traffic Signal Control On Emission Reduction In Urban Road Networks Empowered By Cooperative Vehicle-infrastructure Systems. Applied Energy, 2025;390.
  • Elsevier Sci Ltd, 125 London Wall, London, England (Elsevier - www.elsevier.com; Applied Energy - www.journals.elsevier.com/applied-energy/).
  • Huibo Bi, Beijing University of Technology, College of Metropolitan Transportation, Beijing, People's Republic of China.