发表机构
University of Illinois at Urbana-Champaign（伊利诺伊大学厄巴纳-香槟分校）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
针对分布偏移下群体导航的不确定性，提出自适应椭球共形预测与CVaR调节的拉格朗日策略优化，提升OOD成功率并降低碰撞率，且可部署于物理机器人。
AI 中文摘要
在分布偏移下进行安全的群体导航，需要能够捕捉结构化人体运动预测误差的不确定性表示，以及能够考虑罕见但后果严重的安全目标。现有的不确定性感知方法通常使用各向同性区域来表示预测误差，这些区域可能过于保守，或与方向性运动不确定性对齐不佳。我们提出了一种风险感知导航框架，该框架使用各向异性共形椭球将结构化预测不确定性转化为情节级条件风险价值（CVaR）信号，以调节拉格朗日安全惩罚。具体而言，自适应椭球共形预测（AECP）捕捉方向性预测误差，并在分布偏移下自适应校准不确定性区域，而由此产生的CVaR调节导航策略则使用拉格朗日近端策略优化进行优化。我们在分布内设置和涉及行人运动模式偏移的分布外（OOD）设置下评估了所提出的方法。与最先进的基线相比，我们的方法在保持分布内性能竞争力的同时，在OOD设置下将成功率提高了5.68-7.44个百分点，并将碰撞率降低了5.44-6.64个百分点。我们进一步将训练好的策略无需微调部署到具有机载感知和仅CPU推理的物理机器人上，表明整个流程在物理群体导航中是可行的。
英文摘要
Safe crowd navigation under distribution shift requires uncertainty representations that capture structured human-motion prediction errors and safety objectives that account for rare but consequential failures. Existing uncertainty-aware methods typically represent prediction errors using isotropic regions, which can be either overly conservative or poorly aligned with directional motion uncertainty. We introduce a risk-aware navigation framework that uses anisotropic conformal ellipsoids to translate structured prediction uncertainty into an episode-level conditional value-at-risk (CVaR) signal that regulates the Lagrangian safety penalty. In particular, adaptive ellipsoidal conformal prediction (AECP) captures directional prediction errors and adaptively calibrates uncertainty regions under distribution shift, while the resulting CVaR-regulated navigation policy is optimized using Lagrangian proximal policy optimization. We evaluate our proposed method under both in-distribution settings and out-of-distribution (OOD) settings involving shifts in pedestrian motion patterns. Compared with state-of-the-art baselines, our method maintains competitive in-distribution performance while improving success rates by 5.68-7.44 percentage points and reducing collision rates by 5.44-6.64 percentage points across OOD settings. We further deploy the trained policy without fine-tuning on a physical robot with onboard perception and CPU-only inference, showing that the full pipeline is feasible in physical crowd navigation.
