论文/arXiv 2304.13705·2023-04-23

ACT算法赋能低成本硬件实现双臂精细操作成功率达90%

研究提出Action Chunking with Transformers (ACT)算法，解决低成本机器人双臂精细操作难题。仅凭10分钟遥操作演示数据，该系统在穿扎带、插电池等6项高难度任务中实现80-90%成功率，验证了端到端模仿学习在低精度硬件上的有效性。

AI 整理

• 仿真与训练
• 数据与采集
• 具身模型

本文由 AI 根据公开内容整理，可能有误，请以原文为准：Learning Fine-Grained Bimanual Manipulation with Low-Cost Hardware（arXiv 2304.13705）

核心内容

针对机器人精细操作依赖昂贵硬件与精密校准的行业痛点，arXiv论文提出基于低成本硬件的端到端模仿学习系统。研究团队开发了Action Chunking with Transformers (ACT)算法，通过自定义遥操作接口采集真实演示数据，使机器人在无需高精度传感器前提下掌握复杂技能。实验表明，仅需约10分钟人类演示，系统在开启调味杯、插入电池等6项真实世界任务中达到80-90%成功率，显著降低了具身智能数据采集与部署门槛。

ACT算法核心在于构建动作序列生成模型，有效缓解了传统模仿学习中策略误差累积及人类演示非平稳性问题。该系统采用定制化低成本双臂本体与遥操作接口，直接从RGB视觉反馈中学习接触力协调与闭环控制策略，摆脱了对高端力矩传感器或精确运动学标定的依赖。这种“算法补偿硬件”的技术路径，证明了在消费级硬件上通过高质量小样本数据训练，即可实现工业级精细操作性能，为具身智能规模化落地提供了可复现的低成本范式。

关键看点

• ACT算法机制：采用Transformer架构对动作序列进行分块生成建模，有效抑制长时序模仿学习中的误差累积与非平稳演示噪声
• 小样本高效学习：仅需10分钟真实世界遥操作演示数据，即可在6项高难度精细操作任务中取得80-90%的成功率
• 低成本硬件验证：基于定制低成本双臂本体与简易遥操作接口，无需高精度传感器或精密标定即可实现闭环视觉反馈控制
• 精细操作能力：成功攻克穿扎带、插电池、开透明杯盖等需精密接触力协调的任务，突破廉价硬件性能瓶颈

工程与产业观察

该工作确立了“算法精度补偿硬件公差”的工程范式，证明在双臂精细操作领域，高质量小样本数据配合序列生成模型可替代昂贵传感与标定。这为具身智能从实验室走向低成本量产提供了关键技术支撑，提示产业界应重视遥操作采集效率与动作表征学习，而非单纯堆砌硬件规格。

英文摘要 (Abstract)

Fine manipulation tasks, such as threading cable ties or slotting a battery, are notoriously difficult for robots because they require precision, careful coordination of contact forces, and closed-loop visual feedback. Performing these tasks typically requires high-end robots, accurate sensors, or careful calibration, which can be expensive and difficult to set up. Can learning enable low-cost and imprecise hardware to perform these fine manipulation tasks? We present a low-cost system that performs end-to-end imitation learning directly from real demonstrations, collected with a custom teleoperation interface. Imitation learning, however, presents its own challenges, particularly in high-precision domains: errors in the policy can compound over time, and human demonstrations can be non-stationary. To address these challenges, we develop a simple yet novel algorithm, Action Chunking with Transformers (ACT), which learns a generative model over action sequences. ACT allows the robot to learn 6 difficult tasks in the real world, such as opening a translucent condiment cup and slotting a battery with 80-90% success, with only 10 minutes worth of demonstrations. Project website: https://tonyzhaozh.github.io/aloha/

作者：Tony Z. Zhao, Vikash Kumar, Sergey Levine, Chelsea Finn

来源

本页事实对应的公开页面，以原文为准。

1. arXiv 2304.13705：Learning Fine-Grained Bimanual Manipulation with Low-Cost Hardware访问 2026-10-23