让机器人模仿学习策略跑得比示范快，突破速度瓶颈。
SAIL: Faster-than-Demonstration Execution of Imitation Learning Policies
- 提出SAIL系统，通过动作一致性保持与自适应调速实现高速执行。
- 实测在仿真中提速4倍，真实机器人最高提速3.2倍。
- 适合工业自动化等需高吞吐的机器人任务场景。
离线模仿学习（IL）方法如行为克隆能有效习得复杂的机器人操作技能。然而，现有IL训练策略只能以示范数据中的相同速度执行任务，限制了机器人系统的任务吞吐量，这对工业自动化等应用至关重要。本文提出并形式化了实现视觉-运动策略超示范速度执行的新问题，识别出机器人动力学与状态-动作分布偏移的根本挑战。我们基于核心洞察构建了全栈式SAIL（模仿学习速度适配）系统，包含四个紧密耦合组件：(1) 保持动作一致性的高速动作推理算法，确保高速下的平滑运动；(2) 控制器无关运动目标的高保真跟踪；(3) 基于运动复杂度动态调整执行速度的自适应调速机制；(4) 处理真实系统延迟的动作调度。在12个任务上，涵盖仿真和两个不同的真实机器人平台的实验表明，SAIL在仿真中可实现最高4倍于示范速度的加速，在真实世界中最高达3.2倍加速。
原文摘要 · Abstract (English)
Offline Imitation Learning (IL) methods such as Behavior Cloning are effective at acquiring complex robotic manipulation skills. However, existing IL-trained policies are confined to executing the task at the same speed as shown in demonstration data. This limits the task throughput of a robotic system, a critical requirement for applications such as industrial automation. In this paper, we introduce and formalize the novel problem of enabling faster-than-demonstration execution of visuomotor policies and identify fundamental challenges in robot dynamics and state-action distribution shifts. We instantiate the key insights as SAIL (Speed Adaptation for Imitation Learning), a full-stack system integrating four tightly-connected components: (1) a consistency-preserving action inference algorithm for smooth motion at high speed, (2) high-fidelity tracking of controller-invariant motion targets, (3) adaptive speed modulation that dynamically adjusts execution speed based on motion complexity, and (4) action scheduling to handle real-world system latencies. Experiments on 12 tasks across simulation and two real, distinct robot platforms show that SAIL achieves up to a 4x speedup over demonstration speed in simulation and up to 3.2x speedup in the real world. Additional detail is available at https://nadunranawaka1.github.io/sail-policy
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。