Comfort by Construction: Adaptive, Comfort-Bounded Action Spaces for Learned Driving Policies

Robust control policy learning for autonomous driving requires training environments to be both physically realistic and computationally scalable, properties that existing simulators provide only in isolation. We introduce Sim2Sim2Sim, a framework that bridges high-fidelity vehicle simulation and scalable reinforcement learning by distilling simulator dynamics into a highly parallelizable learned dynamics model. By training control policies purely within this distilled environment and deploying them back into the high-fidelity source simulator, we demonstrate more efficient policy optimization and reliable transfer under challenging dynamics. We further show that predictive accuracy alone does not fully characterize a learned dynamics model's suitability as a reinforcement learning training environment, which should also be assessed by the quality of the policies it enables.

Faithfully evaluating end-to-end driving policies in simulation requires observations that are not merely photo-realistic, but preserve the scene features a policy relies on to make decisions. Existing platforms, however, exhibit a sim-to-real visual gap that corrupts policy perception, undermining their ability to assess a policy's closed-loop decision-making. To this end, we propose DreamStream, a generative, closed-loop simulator that achieves policy-oriented fidelity using a simulator-grounded autoregressive video model. Our video model is distilled from a large pretrained video model via traffic layout guidance, varying visual appearance while preserving policy-relevant features such as scenario layout and the temporal consistency of dynamic objects. We further observe that perceptual metrics like FID misrank how well these features are preserved. To tackle this, we introduce FD$π$, a new multi-representation metric that measures the sim-to-real gap as the Fréchet distance over scene-context features from public E2E policies. Under FD$π$, DreamStream improves over the strongest prior closed-loop simulator by $1.6\times$ on nuScenes and $4.7\times$ on NAVSIM, and induces the least perturbation to policy's perceptual observability. Based on DreamStream, we construct Navhard-CL benchmark, which turns non-reactive real-world benchmark NAVSIM into interactive testing environments with adversarial driving behaviors and weather variations. This benchmark exposes many failure modes of driving policies, such as scorer bias and lack of recovery behaviors, that prior closed-loop benchmarks overlook. Code and data are available at https://github.com/VAIL-UCLA/DreamStream.

End-to-end (E2E) driving policies have progressed rapidly on open-loop (OL) benchmarks, yet OL evaluation cannot reveal whether a policy withstands compounding errors, recovers from failures, or interacts safely with surrounding actors. We introduce NavSafe-$\infty$, a photorealistic closed-loop (CL) benchmark of 280 scenarios spanning 28 event types, each with success and failure criteria defined within a structured traffic-safety taxonomy, which yields category-level capability scores for Traffic Crashes, Vulnerable Road User Crashes, Traffic Violations, and Traffic Incidents. Evaluating 20 E2E policies, we find that OL gains do not reliably transfer to CL safety. Analyzing two common remedies further shows that passive demonstration perturbation helps mainly when CL rollouts stay near its perturbed training states, and that OL reinforcement-learning fine-tuning exhibits reward hacking by trading safety margin for ego progress, which CL feedback amplifies into compounding safety-critical errors. Together, these results demonstrate the blind spot of OL benchmarks indicating CL safety success. The benchmark and an extensible toolbox for customizable event curation and policy diagnosis will be open-sourced and maintained to facilitate future research.