| Qian, Lang, Sun, Peng, Jin, Jiayue, Cheng, Jing, Jiang, Weiwei, Zhou, Wei and Liu, Yang 2026. EFE-TSC: Efficient fuel-economic traffic signal control via hotspot-aware multi-objective learning. IEEE Transactions on Intelligent Transportation Systems 10.1109/TITS.2026.3730824 |
Abstract
Sustainable urban mobility is central to the 15-minute city vision, yet vehicle emissions at signalized intersections remain a critical bottleneck. While Deep Reinforcement Learning (DRL) has advanced traffic signal control (TSC), existing approaches prioritize travel time over fuel economy, and fuel-conscious methods suffer from redundant state representations and rigid decision fusion. To mitigate these impediments, this research puts forward an Efficient Fuel-Economic Traffic Signal Control via hotspot-aware multi-objective learning (EFE-TSC). Specifically, we formulate a traffic hotspot discrimination mechanism to alleviate data redundancy in state representation whilst furnishing guidance for speed control target specification. We additionally devise an attention-based multi-objective DRL framework utilizing cross-attention mechanisms to adaptively compute the importance weighting of each objective according to prevailing traffic circumstances, effectively transcending constraints characteristic of conventional voting-driven approaches. Moreover, we introduce a vehicle speed control module founded upon the Krauss safety distance model, enabling smoother deceleration trajectories that diminish fuel consumption while guaranteeing secure car-following conduct. Empirical results evidence that our EFE-TSC can achieve better fuel consumption reduction whilst preserving favorable traffic efficiency compared to state-of-the-art methods.
| Item Type: | Article |
|---|---|
| Date Type: | Published Online |
| Status: | In Press |
| Schools: | Schools > Computational & Mathematical Sciences Schools > Computer Science & Informatics |
| Publisher: | Institute of Electrical and Electronics Engineers |
| ISSN: | 1524-9050 |
| Last Modified: | 05 Oct 2026 10:30 |
| URI: | https://orca.cardiff.ac.uk/id/eprint/189984 |
Actions (repository staff only)
 |
Edit Item |
Dimensions
Dimensions