arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2606.25985cs.RO
收藏

Action ControlNet: 一种轻量级延迟感知适配器,用于视觉-语言-动作模型中的平滑异步控制

Action ControlNet: A Lightweight Delay-Aware Adapter for Smooth Asynchronous Control in Vision-Language-Action Models

School of Advanced Manufacturing and Robotics, Peking University(北京大学先进制造与机器人学院)

Tiecheng Guo, Meng Guo

AI总结:

提出Action ControlNet,一种轻量级延迟感知适配器,通过将已执行的动作后缀作为残差条件,在不重新训练骨干网络的情况下,减少异步执行中的动作不连续和抖动,提升机器人操控的鲁棒性和平滑性。

摘要:

视觉-语言-动作(VLA)模型在通用机器人操控中展现出强大潜力,但其推理延迟仍是实现稳定高频控制的主要障碍。异步执行通过重叠策略推理与动作执行来缓解这一瓶颈,然而,当机器人继续移动时,下一个动作块仍基于过时的观测进行预测。因此,直接拼接动作块会导致交接不连续、动作抖动以及在接触密集操控中的失败。现有的补救措施通常需要完全重新训练策略或依赖于特定架构的运行时逻辑。本文提出Action ControlNet(ACNet),一种轻量级延迟感知适配器,它将已执行的动作后缀作为残差条件,作用于基本冻结的动作头。ACNet保持预训练骨干网络不变,引入少量可训练参数,并与扩散和流匹配等生成式动作头兼容。在Kinetix、Meta-World MT50以及真实世界的SO-ARM101平台上,ACNet在推理延迟下提升了鲁棒性,并比直接拼接动作块产生更平滑的异步轨迹,同时比完全延迟条件重新训练更轻量。

英文摘要:

Vision-language-action (VLA) models have shown strong potential for general-purpose robot manipulation, but their inference latency remains a major obstacle to stable high-frequency control. Asynchronous execution mitigates this bottleneck by overlapping policy inference with action execution, yet the next action chunk is still predicted from stale observations while the robot continues to move. Direct chunk stitching therefore introduces handoff discontinuities, action jitter, and failures in contact-rich manipulation. Existing remedies typically require either full-policy retraining or architecture-specific runtime logic. This work proposes Action ControlNet (ACNet), a lightweight delay-aware adapter that uses the executed motion suffix as a residual condition for a mostly frozen action head. ACNet leaves the pretrained backbone unchanged, introduces few trainable parameters, and remains compatible with generative action heads such as diffusion and flow matching. On Kinetix, Meta-World MT50, and a real-world SO-ARM101 platform, ACNet improves robustness under inference delay and yields smoother asynchronous trajectories than direct chunk stitching, while remaining more lightweight than full delay-conditioned retraining.

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。

↑