通过VLM规划的原子技能演示进行温和操作策略学习
Gentle Manipulation Policy Learning via Demonstrations from VLM Planned Atomic Skills
AI总结：
本文提出一种结合层次语义分解、强化学习、视觉语言模型与知识蒸馏的框架，将复杂操作任务分解为原子技能并在仿真中训练，通过VLM生成演示并蒸馏为统一策略，实现无需人类演示的长时程温和操作学习与可扩展泛化。
摘要：
长时程、接触丰富的操作任务的自主执行传统上需要大量真实世界数据和专家工程，带来了显著的成本和可扩展性挑战。本文提出了一个新颖的框架，整合了层次化语义分解、强化学习（RL）、视觉语言模型（VLM）和知识蒸馏来克服这些限制。复杂任务被分解为原子技能，每个原语的RL训练策略完全在仿真中完成。关键的是，我们的RL公式引入了显式的力约束，以防止在精细交互过程中损坏物体。VLM执行高层任务分解和技能规划，生成多样化的专家演示。这些演示通过视觉-触觉扩散策略（Visual-Tactile Diffusion Policy）蒸馏为统一策略，用于端到端执行。我们进行了全面的消融研究，探索基于不同VLM的任务规划器，以确定最优的演示生成流程，并系统地比较了用于技能蒸馏的模仿学习算法。大量的仿真实验和物理部署验证了我们的方法能够在无需昂贵人类演示的情况下实现长时程操作策略学习，同时VLM引导的原子技能框架能够实现向多样化任务的可扩展泛化。
英文摘要：
Autonomous execution of long-horizon, contact-rich manipulation tasks traditionally requires extensive real-world data and expert engineering, posing significant cost and scalability challenges. This paper proposes a novel framework integrating hierarchical semantic decomposition, reinforcement learning (RL), visual language models (VLMs), and knowledge distillation to overcome these limitations. Complex tasks are decomposed into atomic skills, with RL-trained policies for each primitive exclusively in simulation. Crucially, our RL formulation incorporates explicit force constraints to prevent object damage during delicate interactions. VLMs perform high-level task decomposition and skill planning, generating diverse expert demonstrations. These are distilled into a unified policy via Visual-Tactile Diffusion Policy for end-to-end execution. We conduct comprehensive ablation studies exploring different VLM-based task planners to identify optimal demonstration generation pipelines, and systematically compare imitation learning algorithms for skill distillation. Extensive simulation experiments and physical deployment validate that our approach achieves policy learning for long-horizon manipulation without costly human demonstrations, while the VLM-guided atomic skill framework enables scalable generalization to diverse tasks.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
