# ROMAN: Reward-Orchestrated Multi-Head Attention Network for Autonomous Driving System Testing Source: [ROMAN: Reward-Orchestrated Multi-Head Attention Network for Autonomous Driving System Testing]() Collection index: [llms.txt]() Data updated: 2026-03-27T05:00:12.960Z Authors: Jianlei Chi, Yuzhen Wu, Jiaxuan Hou, Xiaodong Zhang, Ming Fan, Suhui Sun, Weijun Dai, Bo Li, Jianguo Sun, Jun Sun Published: 2026-02-05T13:09:58.000Z ## Abstract Automated Driving System (ADS) acts as the brain of autonomous vehicles, responsible for their safety and efficiency. Safe deployment requires thorough testing in diverse real-world scenarios and compliance with traffic laws like speed limits, signal obedience, and right-of-way rules. Violations like running red lights or speeding pose severe safety risks. However, current testing approaches face significant challenges: limited ability to generate complex and high-risk law-breaking scenarios, and failing to account for complex interactions involving multiple vehicles and critical situations. To address these challenges, we propose ROMAN, a novel scenario generation approach for ADS testing that combines a multi-head attention network with a traffic law weighting mechanism. ROMAN is designed to generate high-risk violation scenarios to enable more thorough and targeted ADS evaluation. The multi-head attention mechanism models interactions among vehicles, traffic signals, and other factors. The traffic law weighting mechanism implements a workflow that leverages an LLM-based risk weighting module to evaluate violations based on the two dimensions of severity and occurrence. We have evaluated ROMAN by testing the Baidu Apollo ADS within the CARLA simulation platform and conducting extensive experiments to measure its performance. Experimental results demonstrate that ROMAN surpassed state-of-the-art tools ABLE and LawBreaker by achieving 7.91% higher average violation count than ABLE and 55.96% higher than LawBreaker, while also maintaining greater scenario diversity. In addition, only ROMAN successfully generated violation scenarios for every clause of the input traffic laws, enabling it to identify more high-risk violations than existing approaches. ## Recorded benchmark results A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. - [Scenario Generation on Scenario S2](): Scenario Generation; Scenario S2; Inference Time (s): 47. - [Scenario Generation on Scenario S3](): Scenario Generation; Scenario S3; Inference Time (s): 52. - [Scenario Generation on Scenario S1](): Scenario Generation; Scenario S1; Inference Time (s): 44. - [Scenario Generation on Scenario S4](): Scenario Generation; Scenario S4; Inference Time (s): 53. - [Violation Scenario Generation on Scenario S1](): Violation Scenario Generation; Scenario S1; Mean Score: 9.86. - [Violation Scenario Generation on Scenario S2](): Violation Scenario Generation; Scenario S2; Mean Violation: 8.07. - [Violation Scenario Generation on Scenario S3](): Violation Scenario Generation; Scenario S3; Mean Score: 6.41. - [Violation Scenario Generation on Scenario S4](): Violation Scenario Generation; Scenario S4; Mean Violation: 7.15. ## Original sources and implementation [Original paper]() [PDF]()