FREA: Feasibility-Guided Generation of Safety-Critical Scenarios with Reasonable Adversariality

Keyu Chen, Yuheng Lei, Hao Cheng, Haoran Wu, Wenchao Sun, Sifa Zheng
Proceedings of The 8th Conference on Robot Learning, PMLR 270:566-586, 2025.

Abstract

Generating safety-critical scenarios, which are essential yet difficult to collect at scale, offers an effective method to evaluate the robustness of autonomous vehicles (AVs). Existing methods focus on optimizing adversariality while preserving the naturalness of scenarios, aiming to achieve a balance through data-driven approaches. However, without an appropriate upper bound for adversariality, the scenarios might exhibit excessive adversariality, potentially leading to unavoidable collisions. In this paper, we introduce FREA, a novel safety-critical scenarios generation method that incorporates the Largest Feasible Region (LFR) of AV as guidance to ensure the reasonableness of the adversarial scenarios. Concretely, FREA initially pre-calculates the LFR of AV from offline datasets. Subsequently, it learns a reasonable adversarial policy that controls critical background vehicles (CBVs) in the scene to generate adversarial yet AV-feasible scenarios by maximizing a novel feasibility-dependent objective function. Extensive experiments illustrate that FREA can effectively generate safety-critical scenarios, yielding considerable near-miss events while ensuring AV’s feasibility. Generalization analysis also confirms the robustness of FREA in AV testing across various surrogate AV methods and traffic environments.

Cite this Paper

BibTeX
@InProceedings{pmlr-v270-chen25b, title = {FREA: Feasibility-Guided Generation of Safety-Critical Scenarios with Reasonable Adversariality}, author = {Chen, Keyu and Lei, Yuheng and Cheng, Hao and Wu, Haoran and Sun, Wenchao and Zheng, Sifa}, booktitle = {Proceedings of The 8th Conference on Robot Learning}, pages = {566--586}, year = {2025}, editor = {Agrawal, Pulkit and Kroemer, Oliver and Burgard, Wolfram}, volume = {270}, series = {Proceedings of Machine Learning Research}, month = {06--09 Nov}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v270/main/assets/chen25b/chen25b.pdf}, url = {https://proceedings.mlr.press/v270/chen25b.html}, abstract = {Generating safety-critical scenarios, which are essential yet difficult to collect at scale, offers an effective method to evaluate the robustness of autonomous vehicles (AVs). Existing methods focus on optimizing adversariality while preserving the naturalness of scenarios, aiming to achieve a balance through data-driven approaches. However, without an appropriate upper bound for adversariality, the scenarios might exhibit excessive adversariality, potentially leading to unavoidable collisions. In this paper, we introduce FREA, a novel safety-critical scenarios generation method that incorporates the Largest Feasible Region (LFR) of AV as guidance to ensure the reasonableness of the adversarial scenarios. Concretely, FREA initially pre-calculates the LFR of AV from offline datasets. Subsequently, it learns a reasonable adversarial policy that controls critical background vehicles (CBVs) in the scene to generate adversarial yet AV-feasible scenarios by maximizing a novel feasibility-dependent objective function. Extensive experiments illustrate that FREA can effectively generate safety-critical scenarios, yielding considerable near-miss events while ensuring AV’s feasibility. Generalization analysis also confirms the robustness of FREA in AV testing across various surrogate AV methods and traffic environments.} }
Endnote
%0 Conference Paper %T FREA: Feasibility-Guided Generation of Safety-Critical Scenarios with Reasonable Adversariality %A Keyu Chen %A Yuheng Lei %A Hao Cheng %A Haoran Wu %A Wenchao Sun %A Sifa Zheng %B Proceedings of The 8th Conference on Robot Learning %C Proceedings of Machine Learning Research %D 2025 %E Pulkit Agrawal %E Oliver Kroemer %E Wolfram Burgard %F pmlr-v270-chen25b %I PMLR %P 566--586 %U https://proceedings.mlr.press/v270/chen25b.html %V 270 %X Generating safety-critical scenarios, which are essential yet difficult to collect at scale, offers an effective method to evaluate the robustness of autonomous vehicles (AVs). Existing methods focus on optimizing adversariality while preserving the naturalness of scenarios, aiming to achieve a balance through data-driven approaches. However, without an appropriate upper bound for adversariality, the scenarios might exhibit excessive adversariality, potentially leading to unavoidable collisions. In this paper, we introduce FREA, a novel safety-critical scenarios generation method that incorporates the Largest Feasible Region (LFR) of AV as guidance to ensure the reasonableness of the adversarial scenarios. Concretely, FREA initially pre-calculates the LFR of AV from offline datasets. Subsequently, it learns a reasonable adversarial policy that controls critical background vehicles (CBVs) in the scene to generate adversarial yet AV-feasible scenarios by maximizing a novel feasibility-dependent objective function. Extensive experiments illustrate that FREA can effectively generate safety-critical scenarios, yielding considerable near-miss events while ensuring AV’s feasibility. Generalization analysis also confirms the robustness of FREA in AV testing across various surrogate AV methods and traffic environments.
APA
Chen, K., Lei, Y., Cheng, H., Wu, H., Sun, W. & Zheng, S.. (2025). FREA: Feasibility-Guided Generation of Safety-Critical Scenarios with Reasonable Adversariality. Proceedings of The 8th Conference on Robot Learning, in Proceedings of Machine Learning Research 270:566-586 Available from https://proceedings.mlr.press/v270/chen25b.html.

Related Material