Formally Constrained Reinforcement Learning for Traffic Signal Control at Intersections
Bibliographic record
Abstract
Ensuring the safety of autonomous systems is paramount for their successful integration into real-world transportation networks. Autonomous Vehicles and Machine Learning-driven traffic management systems have the potential to enhance efficiency and mobility. However, their deployment presents significant safety challenges, particularly in managing interactions between autonomous systems, human-driven vehicles, and pedestrians. This paper addresses these challenges by focusing on Reinforcement Learning (RL)-based traffic signal control. It analyzes traffic safety using Time-To-Collision to identify potential traffic conflicts between vehicle interactions that could jeopardize safety. We propose a novel approach to mitigate these conflicts by guiding the learning process under a formally checked safety constraint. This approach leverages Satisfiability Modulo Theories as a formal method for rigorous verification to ensure that the RL agent's decisions remain within safe operational boundaries, even in dynamic and unpredictable traffic conditions. Additionally, our approach incorporates dynamic speed adjustment mechanisms to address scenarios in which safety constraints are violated. Through traffic simulations, we evaluate the effectiveness of our approach in achieving a balance between traffic signal optimization and traffic safety by ensuring the safe operation of RL under a safety constraint and demonstrate that this adaptive speed control strategy reduces both the frequency and severity of traffic conflicts.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.002 | 0.011 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.002 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.002 |
| Insufficient payload (model declined to judge) | 0.002 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".