Scholarlink

<p>Traffic congestion is a major problem in urban areas due to urbanisation and increased vehicle usage. Conventional traffic signal control strategies, such as fixed-time and adaptive approaches, are unable to handle dynamic and unpredictable traffic conditions. To overcome these limitations, reinforcement learning (RL)- based techniques have emerged as promising solutions for adaptive traffic signal control using real-time traffic information. However, most existing review studies analyse traditional optimisation techniques and RL-based approaches separately without providing a unified comparative framework. This study proposes a comprehensive review of traditional optimisation methods, emerging intelligent techniques, and RL-based traffic signal control approaches within a unified analytical framework. Practical challenges of RL-based traffic signal control, including scalability, interpretability, deployment complexity, and dependence on large-scale data, are also discussed in the study. The reviewed studies indicate that RL-based approaches provide greater flexibility and improved performance than the traditional methods. Finally, the study identifies current research gaps and future research directions for developing efficient and reliable intelligent transportation systems.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A Systematic Review of Traditional and Reinforcement Learning-Based Traffic Signal Control Methods

  • 期刊论文

  • 发表日期：

    2026年6月16日

查看全文
  • R. Manikandan,
  • J. Arun Pandian

摘要

Traffic congestion is a major problem in urban areas due to urbanisation and increased vehicle usage. Conventional traffic signal control strategies, such as fixed-time and adaptive approaches, are unable to handle dynamic and unpredictable traffic conditions. To overcome these limitations, reinforcement learning (RL)- based techniques have emerged as promising solutions for adaptive traffic signal control using real-time traffic information. However, most existing review studies analyse traditional optimisation techniques and RL-based approaches separately without providing a unified comparative framework. This study proposes a comprehensive review of traditional optimisation methods, emerging intelligent techniques, and RL-based traffic signal control approaches within a unified analytical framework. Practical challenges of RL-based traffic signal control, including scalability, interpretability, deployment complexity, and dependence on large-scale data, are also discussed in the study. The reviewed studies indicate that RL-based approaches provide greater flexibility and improved performance than the traditional methods. Finally, the study identifies current research gaps and future research directions for developing efficient and reliable intelligent transportation systems.