RiskFlow: Fast and Faithful Safety-Critical Traffic Scenario Generation
RiskFlow: Fast and Faithful Safety-Critical Traffic Scenario Generation
This research presents RiskFlow, a closed-loop framework for generating safety-critical traffic scenarios. It reformulates future behavior generation as one-step flow transport in the action space during training to avoid accumulated errors from iterative denoising. At test time, it uses time-to-collision-based ranking to select high-risk agents and an adversarial objective to refine their actions, with vehicle-dynamics rollout ensuring feasible trajectories. Experiments on the nuScenes dataset show RiskFlow improves realism while maintaining safety-critical capability and significantly reduces inference time compared to baselines.
The paper introduces RiskFlow, a framework for generating safety-critical traffic scenarios using one-step flow transport to reduce computational cost and motion artifacts compared to iterative diffusion methods.
Source: arXiv Robotics — research abstracts · Read original article ↗
Loading article text…
What the source reports
Publisher-reported claims, with original evidence. These results have not been independently verified by RoboSignal.
- paper: not_reported
Source excerpts and review record
Automatically extracted; no manual editorial approval recorded.
ision-based critical-agent ranking selects high-risk agents and an adversarial interaction objective directly refines their output actions. Vehicle-dynamics rollout and map-aware regularization then translate guided actions into feasible trajectories while discouraging off-road behavior. By intervening on selected agents in the output space, RiskFlow avoids repeated corrections of intermediate denoising states, there
Open source S4
by mitigating accumulated guidance errors and accelerating closed-loop scenario generation. Experiments on the nuScenes dataset under closed-loop evaluation show that RiskFlow achieves a strong adversariality-realism trade-off across multi-agent and long-horizon settings. Compared with representative baselines, RiskFlow consistently improves realism while maintaining competitive safety-critical generation capability,
Open source S5
and substantially reduces inference time for evaluation.
Open source S6
RiskFlow: Fast and Faithful Safety-Critical Traffic Scenario Generation
Open source S1
arXiv:2606.06423v2 Announce Type: replace Abstract: Safety-critical traffic scenario generation is essential for evaluating autonomous driving systems under rare but high-risk interactions. Existing diffusion-based methods offer strong controllability in closed-loop generation, but their iterative denoising process is computationally expensive and may accumulate sampling and guidance errors over long rollouts, causin
Open source S2
g unrealistic motion artifacts such as abnormal acceleration and off-road behavior. To address these issues, we propose RiskFlow, a closed-loop safety-critical traffic generation framework that reformulates future behavior generation as one-step flow transport in the action space during training. At test time, RiskFlow generates safety-critical interactions through a selective guidance procedure in which time-to-coll
Open source S3
Source：arXiv Robotics — research abstracts · arxiv.org