As a government-regulated public service, traffic signal control (TSC) requires reliable and transparent decision-making. However, existing deep reinforcement learning (DRL) methods, despite improvements in control accuracy, still lack explainability and generalisation, severely limiting their applicability in real-world environments. To address the challenges above, this paper proposes GenEx-TSC, a generalisable and explainable TSC method that integrates deep reinforcement learning with large language models (LLMs). First, starting from vehicle-level states, we train a DRL agent incorporating intersection physical heterogeneity and neighbourhood information, which lays the evaluation foundation for constructing a high-quality LLM dataset. Subsequently, the LLM agent is optimised through a two-stage training mechanism. In the distillation stage, a lightweight LLM agent is trained using the reasoning trajectories of a larger-scale LLM agent, inheriting its semantic understanding and decision-generation capabilities and in the alignment stage, the DRL evaluation network is employed to calibrate the outputs of the distilled LLM agent, ensuring that the generated cycle-level signal timing strategies are both efficient and interpretable. We synthesise 10 intersection networks with different physical attributes in SUMO and set traffic flows of varying scales. Experimental results across diverse traffic environments demonstrate that the proposed GenEx-TSC exhibits clear advantages over traditional methods, mainstream DRL methods and LLM baselines in terms of control accuracy, generalisation and explainability.
Funding
This work was support in part by the National Natural Science Foundation of China under (Grant No. 62501094); and in part by the Natural Science Foundation of Chongqing under (Grant Nos. CSTB2025NSCQ-LZX0152, CSTB2024NSCQ-LZX0134 and CSTB2025NSCQ-LZX0052).
Conflicts of Interest
The authors declare no conflicts of interest.
Data Availability Statement
Data available on request from the authors.
| [1] |
S. Xu, C. Sun and N. Liu, “Road Congestion and Air Pollution Analysis of Spatial and Temporal Congestion Effects,” Science of the Total Environment 945 (2024): 173896, https://doi.org/10.1016/j.scitotenv.2024.173896.
|
| [2] |
J. Yu, L. Wang and X. Gong, “Study on the Status Evaluation of Urban Road Intersections Traffic Congestion Base on Ahp Topsis Modal,” Procedia Social and Behavioral Sciences 96 (2013): 609-616, https://doi.org/10.1016/j.sbspro.2013.08.071.
|
| [3] |
M. Elassy, M. Al Hattab, M. Takruri and S. Badawi, “Intelligent Transportation Systems for Sustainable Smart Cities,” Transport Engineer 16 (2024): 100252, https://doi.org/10.1016/j.treng.2024.100252.
|
| [4] |
H. Zhao, C. Dong, J. Cao and Q. Chen, “A Survey on Deep Reinforcement Learning Approaches for Traffic Signal Control,” Engineering Applications of Artificial Intelligence 133 (2024): 108100, https://doi.org/10.1016/j.engappai.2024.108100.
|
| [5] |
L. Li, Y. Lv and F. Y. Wang, “Traffic Signal Timing via Deep Reinforcement Learning,” IEEE/CAA Journal of Automatica Sinica 3, no. 3 (2016): 247-254, https://doi.org/10.1109/jas.2016.7508798.
|
| [6] |
D. Ma, B. Zhou, X. Song and H. Dai, “A Deep Reinforcement Learning Approach to Traffic Signal Control With Temporal Traffic Pattern Mining,” IEEE Transactions on Intelligent Transportation Systems 23, no. 8 (2021): 11789-11800, https://doi.org/10.1109/tits.2021.3107258.
|
| [7] |
K. F. Chu, A. Y. Lam and V. O. Li, “Traffic Signal Control Using End to End Off Policy Deep Reinforcement Learning,” IEEE Transactions on Intelligent Transportation Systems 23, no. 7 (2021): 7184-7195, https://doi.org/10.1109/tits.2021.3067057.
|
| [8] |
Q. Wu, J. Wu, J. Shen, et al., Distributed Agent Based Deep Reinforcement Learning for Large Scale Traffic Signal Control,” Knowledge Based Systems 241 (2022): 108304, https://doi.org/10.1016/j.knosys.2022.108304.
|
| [9] |
B. Liu, K. Su, E. Wang, et al., Matlit: Mat Based Cooperative Reinforcement Learning for Urban Traffic Signal Control,” IEEE Transactions on Intelligent Transportation Systems 26, no. 5 (2025): 6675-6689, https://doi.org/10.1109/tits.2025.3536541.
|
| [10] |
H. Gu, S. Wang, D. Jia, et al., Communication Strategy on Macro and Micro Traffic State in Cooperative Deep Reinforcement Learning for Regional Traffic Signal Control,” IEEE Transactions on Intelligent Transportation Systems 26, no. 8 (2025): 12183-12196, https://doi.org/10.1109/tits.2025.3556931.
|
| [11] |
H. Huang, Z. Hu, Z. Lu and X. Wen, “Network Scale Traffic Signal Control via Multiagent Reinforcement Learning With Deep Spatiotemporal Attentive Network,” IEEE Transactions on Cybernetics 53, no. 1 (2021): 262-274, https://doi.org/10.1109/tcyb.2021.3087228.
|
| [12] |
Y. Bie, Y. Ji and D. Ma, “Multi Agent Deep Reinforcement Learning Collaborative Traffic Signal Control Method Considering Intersection Heterogeneity,” Transportation Research Part C: Emerging Technologies 164 (2024): 104663, https://doi.org/10.1016/j.trc.2024.104663.
|
| [13] |
C.Mingyang, H.Huang, X.Qing, et al., Survey of Intelligent and Connected Vehicle Technologies: Architectures, Functions and Applications,” Journal of Tsinghua University 62, no. 3 (2022): 493-508, https://doi.org/10.16511/j.cnki.qhdxxb.2021.26.026.
|
| [14] |
Y. Lu, W. Wang, R. Bai, et al., Hyper Relational Interaction Modeling in Multi Modal Trajectory Prediction for Intelligent Connected Vehicles in Smart Cites,” Information Fusion 114 (2025): 102682, https://doi.org/10.1016/j.inffus.2024.102682.
|
| [15] |
Z. Zhai, R. Hao, B. Cui and S. Wang, “Hgat and Multi Agent rl Based Method for Multi Intersection Traffic Signal Control,” IEEE Transactions on Intelligent Transportation Systems 26, no. 5 (2025): 6848-6864, https://doi.org/10.1109/tits.2025.3526223.
|
| [16] |
H. Huang, H. Qi, Z. Hu, et al., Inductive Meta Deep Reinforcement Learning for Traffic Signal Control Considering Heterogeneity in Traffic Environments,” IEEE Transactions on Vehicular Technology (2025).
|
| [17] |
J. Xing, T. Nagata, K. Chen, X. Zou, E. Neftci and J. L. Krichmar, “Domain Adaptation in Reinforcement Learning via Latent Unified State Representation,” in Proceedings of the AAAI Conference on Artificial Intelligence, Vol. 35 (2021), 10452-10459, https://doi.org/10.1609/aaai.v35i12.17251.
|
| [18] |
Y. Xiong, Z. Hu, Y. Huang, et al., Xrl Bench: A Benchmark for Evaluating and Comparing Explainable Reinforcement Learning Techniques,” in Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, (2024), 6073-6082.
|
| [19] |
G. Kim, J. Kang and K. Sohn, “A Meta Reinforcement Learning Algorithm for Traffic Signal Control to Automatically Switch Different Reward Functions According to the Saturation Level of Traffic Flows,” Computer Aided Civil and Infrastructure Engineering 38, no. 6 (2023): 779-798, https://doi.org/10.1111/mice.12924.
|
| [20] |
X. Zang, H. Yao, G. Zheng, N. Xu, K. Xu and Z. Li, “Metalight: Value Based Meta Reinforcement Learning for Traffic Signal Control,” in Proceedings of the AAAI Conference on Artificial Intelligence, (2020), 1153-1160, https://doi.org/10.1609/aaai.v34i01.5467.
|
| [21] |
Y. Gu, K. Zhang, Q. Liu, W. Gao, L. Li and J. Zhou, “pi Light: Programmatic Interpretable Reinforcement Learning for Resource Limited Traffic Signal Control,” in Proceedings of the AAAI Conference on Artificial Intelligence, Vol. 38, (2024), 21107-21115.
|
| [22] |
Y. Zhu, X. Yin and C. Chen, “Extracting Decision Tree From Trained Deep Reinforcement Learning in Traffic Signal Control,” IEEE Transactions on Computational Social Systems 10, no. 4 (2022): 1997-2007, https://doi.org/10.1109/tcss.2022.3225362.
|
| [23] |
H. Naveed, A. U. Khan, S. Qiu, et al., A Comprehensive Overview of Large Language Models,” ACM Transactions on Intelligent Systems and Technology 16, no. 5 (2025): 1-72, https://doi.org/10.1145/3744746.
|
| [24] |
S. Wandelt, C. Zheng, S. Wang, Y. Liu and X. Sun, “Large Language Models for Intelligent Transportation: A Review of the State of the Art and Challenges,” Applied Sciences 14, no. 17 (2024): 7455, https://doi.org/10.3390/app14177455.
|
| [25] |
L. Da, M. Gao, H. Mei and H. Wei, “Prompt to Transfer: Sim To Real Transfer for Traffic Signal Control With Prompt Learning,” in Proceedings of the AAAI Conference on Artificial Intelligence, Vol. 38, (2024), 82-90, https://doi.org/10.1609/aaai.v38i1.27758.
|
| [26] |
M. Wang, A. Pang, Y. Kan, M. O. Pun, C. S. Chen and B. Huang, “Llm Assisted Light: Leveraging Large Language Model Capabilities for Human Mimetic Traffic Signal Control in Complex Urban Environments,” preprint arXiv:2403.08337 (2024).
|
| [27] |
M. Movahedi and J. Choi, “The Crossroads of Llm and Traffic Control: A Study on Large Language Models in Adaptive Traffic Signal Control,” IEEE Transactions on Intelligent Transportation Systems 26, no. 2 (2025): 1701-1716, https://doi.org/10.1109/tits.2024.3498735.
|
| [28] |
A. Pang, M. Wang, M. O. Pun, C. S. Chen and X. Xiong, “Illm tsc: Integration Reinforcement Learning and Large Language Model for Traffic Signal Control Policy Improvement,” preprint arXiv:2407.06025 (2024).
|
| [29] |
S. Lai, Z. Xu, W. Zhang, H. Liu and H. Xiong, “Llmlight: Large Language Models as Traffic Signal Control Agents,” in Proceedings of the 31st ACM SIGKDD Conference on Knowledge Discovery and Data Mining V, Vol. 1, (2025), 2335-2346.
|
| [30] |
H. Huang, Z. Hu, M. Li, Z. Lu and X. Wen, “Cooperative Optimization of Traffic Signals and Vehicle Speed Using a Novel Multi Agent Deep Reinforcement Learning,” IEEE Transactions on Vehicular Technology 73, no. 6 (2024): 7785-7798, https://doi.org/10.1109/tvt.2024.3359282.
|
| [31] |
S. Fujimoto, H. Hoof and D. Meger, “Addressing Function Approximation Error in Actor Critic Methods,” in International Conference on Machine Learning (PMLR, 2018), 1587-1596.
|
| [32] |
H. Wei, C. Chen, G. Zheng, et al., Presslight: Learning Max Pressure Control to Coordinate Traffic Signals in Arterial Network,” in Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, (2019), 1290-1298.
|
| [33] |
P. Velickovic, G. Cucurull, A. Casanova, et al., Graph Attention Networks,” in Proceedings of the 6th International Conference on Learning Representations (ICLR, 2018), https://doi.org/10.17863/CAM.48429.
|
| [34] |
B. Ji, H. Liu, M. Du and S. K. Ng, “Chain of Thought Improves Text Generation With Citations in Large Language Models,” in Proceedings of the AAAI Conference on Artificial Intelligence, Vol. 38, (2024), 18345-18353, https://doi.org/10.1609/aaai.v38i16.29794.
|
| [35] |
M. Štrupl, F. Faccio, D. R. Ashley, R. K. Srivastava and J. Schmidhuber, “Reward Weighted Regression Converges to a Global Optimum,” in Proceedings of the AAAI Conference on Artificial Intelligence, Vol. 36, (2022), 8361-8369.
|
| [36] |
P. A. Lopez, M. Behrisch, L. Bieker Walz, et al., Microscopic Traffic Simulation Using Sumo,” in 2018 21st International Conference on Intelligent Transportation Systems (ITSC) (IEEE, 2018), 2575-2582.
|
| [37] |
A. Wegener, M. Piórkowski, M. Raya, H. Hellbrück, S. Fischer and J. P. Hubaux, “Traci: An Interface for Coupling Road Traffic and Network Simulators,” in Proceedings of the 11th Communications and Networking Simulation Symposium, (2008), 155-163.
|
| [38] |
P. Koonce, “‘Traffic Signal Timing Manual,’ United States,” Federal Highway Administration, Tech. Rep. (2008).
|
| [39] |
P. Varaiya, “Max Pressure Control of a Network of Signalized Intersections,” Transportation Research Part C: Emerging Technologies 36 (2013): 177-195, https://doi.org/10.1016/j.trc.2013.08.014.
|
| [40] |
H. Wei, N. Xu, H. Zhang, et al., Colight: Learning Network Level Cooperation for Traffic Signal Control,” in Proceedings of the 28th ACM International Conference on Information and Knowledge Management, (2019), 1913-1922.
|
| [41] |
A. Liu, B. Feng, B. Xue, et al., Deepseek v3 Technical Report,” preprint arXiv:2412.19437 (2024).
|
| [42] |
E. J. Hu, Y. Shen, P. Wallis, et al., Lora: Low Rank Adaptation of Large Language Models,” International Conference on Learning Representations, Vol. 1, (2022), 3.
|
PDF
(2220KB)
