aiqbench analysis
This research addresses a critical gap in autonomous driving validation by generating high-fidelity safety-critical scenarios using a Transformer-enhanced reinforcement learning framework. The proposed method successfully bridges the divide between sparse real-world accident data and limited simulation capabilities, producing over 198,000 episodes that pass Turing tests for behavioral realism. This advancement significantly lowers the barrier for rigorous testing of vehicle-pedestrian interactions without relying on rare physical incidents.
Generating Realistic Safety-Critical Scenarios for Vehicle-Pedestrian Interactions
arXiv:2605.17229v2 Announce Type: replace Abstract: Automated driving system deployment requires rigorous validation across safety-critical vehicle-pedestrian interactions, yet real-world datasets rarely capture high-risk scenarios while simulation platforms lack realistic behavior. In response, this study proposes a three-stage framework that combines real-world grounding with adaptive simulation to generate behaviorally realistic safety-critical scenarios at scale. Stage 1 pre-trains multi-agent state-space Transformer-enhanced DDPG (MA-SST-DDPG) agents on real-world safety-critical data to learn human-like interactive evasive behaviors through data-driven learning. Stage 2 deploys pre-trained multi-agents in CARLA for online reinforcement learning to generalize across diverse scenarios, integrating real-world knowledge with simulation experience to produce a refined MA-SST-DDPG model. Stage 3 uses CARLA with the refined model to generate over 198,000 high-resolution interaction episodes from eight intersection scenarios, culminating in the Vehicle-Pedestrian Safety-Critical Interaction (VPSCI) dataset. The Refined MA-SST-DDPG model outperformed baseline methods in reproducing realistic evasive behaviors, achieving the lowest trajectory errors (ADE = 0.072 m, FDE = 0.142 m). Statistical comparison confirmed distributional equivalence between the generated and real-world data in both conflict severity and behavioral response. A Turing test confirmed that the three-stage framework generated evasive behaviors were indistinguishable from real-world interactions. These results demonstrate the framework's effectiveness in producing high-fidelity safety-critical data, offering valuable sources for the development of ADS and simulation-based safety evaluations.
Abstract:Automated driving system deployment requires rigorous validation across safety-critical vehicle-pedestrian interactions, yet real-world datasets rarely capture high-risk scenarios while simulation platforms lack realistic behavior. In response, this study proposes a three-stage framework that combines real-world grounding with adaptive simulation to generate behaviorally realistic safety-critical scenarios at scale. Stage 1 pre-trains multi-agent state-space Transformer-enhanced DDPG (MA-SST-DDPG) agents on real-world safety-critical data to learn human-like interactive evasive behaviors through data-driven learning. Stage 2 deploys pre-trained multi-agents in CARLA for online reinforcement learning to generalize across diverse scenarios, integrating real-world knowledge with simulation experience to produce a refined MA-SST-DDPG model. Stage 3 uses CARLA with the refined model to generate over 198,000 high-resolution interaction episodes from eight intersection scenarios, culminating in the Vehicle-Pedestrian Safety-Critical Interaction (VPSCI) dataset. The Refined MA-SST-DDPG model outperformed baseline methods in reproducing realistic evasive behaviors, achieving the lowest trajectory errors (ADE = 0.072 m, FDE = 0.142 m). Statistical comparison confirmed distributional equivalence between the generated and real-world data in both conflict severity and behavioral response. A Turing test confirmed that the three-stage framework generated evasive behaviors were indistinguishable from real-world interactions. These results demonstrate the framework's effectiveness in producing high-fidelity safety-critical data, offering valuable sources for the development of ADS and simulation-based safety evaluations.
Robotics (cs.RO); Systems and Control (eess.SY)
I.2.11; I.2.6; I.6.3; I.2.9
arXiv:2605.17229 [cs.RO]
(or arXiv:2605.17229v2 [cs.RO] for this version)
https://doi.org/10.48550/arXiv.2605.17229 Focus to learn more arXiv-issued DOI via DataCite
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 305 words
Related coverage
Controllable and Photorealistic Pedestrian Risky Motion Generation for End-to-End Driving Safety Evaluation
arXiv · Oct 6, 2026
One Diffusion Model, Two Roles: Guided Trajectory Planning and Safety-Critical Scenario Generation in Closed-Loop Simulation
arXiv · Sep 7, 2026
End-to-end Conditional Diffusion for Realistic and Controllable Visual Traffic Scenario Generation
arXiv · Jul 22, 2026
LLM-Guided Transformation of Non-Critical Driving Scenes into Safety-Critical Scenarios Using Augmented Reality
arXiv · Sep 18, 2026