Towards fair lights: a multi-agent masked deep reinforcement learning for efficient corridor-level traffic signal control

Urban traffic management is gradually shifting from a focus on “efficiency first” to a more balanced approach that considers both efficiency and fairness. In response to the growing need for equitable multimodal mobility, Dr. Xiaocai Zhang and Dr. Neema Nassir from the University of Melbourne, Australia, proposed a human-centric traffic signal control method that balances fairness and efficiency—M2SAC (Multi-agent Masked Soft-Actor-Critic)—aimed at achieving coordinated intelligent control across multiple intersections at the corridor level.

They published their study on 11 August 2025, in Communications in Transportation Research.

With the rapid advancement of AI technologies in transportation, Deep Reinforcement Learning (DRL)-based Adaptive Traffic Signal Control (ATSC) has become a research hotspot. However, most existing methods center on motor vehicle flow, often overlooking the needs of vulnerable road users such as pedestrians and public transit passengers, and struggle to balance fairness and efficiency.

To address the increasingly complex multimodal urban traffic systems, the research team has proposed a more human-oriented intelligent signal control framework, shifting the paradigm from "vehicle-priority" to "equal treatment for all road users". The proposed M2SAC framework incorporates several key innovations:

1. Phase Masking Mechanism: For the first time, a phase “masking” mechanism is introduced, allowing the system to dynamically skip certain signal phases based on traffic conditions, thus enhancing exploration and avoiding local optima.
2. Hybrid Action Space Modeling: Combines Gaussian-distributed outputs for phase masking decisions with categorical-distributed outputs for green light timing ratios, enabling joint optimization of phase order and duration.
3. Fairness-Oriented Objective Function: Uses the “minimum number of people affected” as the core reward signal, comprehensively considering vehicle passengers, transit users, and pedestrians, thereby enhancing the human-centered and fairness attributes of signal timing.

The study is based on a real-world corridor network on Elgin Street in Melbourne, Australia, selecting three signalized intersections to build a realistic simulation environment using the PTV VISSIM platform. Five representative traffic scenarios were designed. In each scenario, M2SAC demonstrated outstanding control performance. Compared to fixed-timing control methods (such as the Webster algorithm and green wave coordination) and several mainstream DRL approaches (e.g., MADDPG, MADDQN, and MAA2C), M2SAC reduced the average number of affected individuals by 5.17% to 34.35%, especially excelling in mixed traffic scenarios involving both vehicles and pedestrians. Moreover, the introduction of the masking mechanism provided an additional 3%–7% improvement across multiple scenarios, verifying the effectiveness of its design.

This study is the first to introduce the concept of "traffic fairness" into deep reinforcement learning-based traffic signal control systems, breaking away from the conventional vehicle-centric mindset and addressing the travel experience of different road users. It offers a practical and intelligent solution toward a more human-centered and equitable traffic system.

The above research is published in Communications in Transportation Research (COMMTR), which is a fully open access journal co-published by Tsinghua University Press and Elsevier. COMMTR publishes peer-reviewed high-quality research representing important advances of significance to emerging transport systems. COMMTR is also among the first transportation journals to make the Replication Package mandatory to facilitate researchers, practitioners, and the general public in understanding and advancing existing knowledge. At its discretion, Tsinghua University Press will pay the open access fee for all published papers in 2025.

About Communications in Transportation Research

Communications in Transportation Research was launched in 2021, with academic support provided by Tsinghua University and China Intelligent Transportation Systems Association. The Editors-in-Chief are Professor Xiaobo Qu, a member of the Academia Europaea from Tsinghua University and Professor Shuai’an Wang from Hong Kong Polytechnic University. The journal mainly publishes high-quality, original research and review articles that are of significant importance to emerging transportation systems, aiming to serve as an international platform for showcasing and exchanging innovative achievements in transportation and related fields, fostering academic exchange and development between China and the global community.

It has been indexed in SCIE, SSCI, Ei Compendex, Scopus, CSTPCD, CSCD, OAJ, DOAJ, TRID and other databases. It was selected as Q1 Top Journal in the Engineering and Technology category of the Chinese Academy of Sciences (CAS) Journal Ranking List. In 2022, it was selected as a High-Starting-Point new journal project of the “China Science and Technology Journal Excellence Action Plan”. In 2024, it was selected as the Support the Development Project of “High-Level International Scientific and Technological Journals”. The same year, it was also chosen as an English Journal Tier Project of the “China Science and Technology Journal Excellence Action Plan Phase Ⅱ”. In 2024, it received the first impact factor (2023 IF) of 12.5, ranking Top1 (1/58, Q1) among all journals in “TRANSPORTATION” category. In 2025, its 2024 IF was announced as 14.5, maintaining the Top 1 position (1/61, Q1) in the same category. Tsinghua University Press will cover the open access fee for all published papers in 2025.