arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2603.25399cs.CVcs.RO
收藏

LaMP: 以3D场景流作为潜在运动先验的视觉-语言-动作策略学习

LaMP: Learning Vision-Language-Action Policy with 3D Scene Flow as Latent Motion Prior

Southeast University(东南大学);School of Artificial Intelligence, Shanghai Jiao Tong University(上海交通大学人工智能学院);Zhejiang University(浙江大学);Beihang University(北京航空航天大学);Shanghai Innovation Institute(上海创新研究院)

Xinkai Wang, Chenyi Wang, Yifu Xu, Mingzhe Ye, Fucheng Zhang, Jialin Tian, Xinyu Zhan, Lifeng Zhu, Cewu Lu, Lixin Yang

AI总结:

提出LaMP框架,通过门控融合将3D场景流作为潜在运动先验,对齐运动专家与动作专家,提升VLA策略在复杂空间交互中的鲁棒性,在LIBERO等基准上取得最高平均成功率。

摘要:

我们介绍了\textbf{LaMP},一个双专家视觉-语言-动作框架,将密集3D场景流嵌入为机器人操作的潜在运动先验。现有的VLA模型直接从2D语义视觉特征回归动作,迫使它们学习复杂的3D物理交互。这种隐式学习策略在陌生空间变化下性能下降。LaMP通过门控融合将对齐流匹配的\textit{运动专家}与预测策略的\textit{动作专家}来解决这一限制。具体地,运动专家生成一步部分去噪的3D场景流,其隐藏状态条件化动作专家,无需完整的多步去噪。我们在LIBERO、LIBERO-Plus和SimplerEnv-WidowX模拟基准以及真实世界任务上评估LaMP。LaMP在LIBERO、LIBERO-Plus和SimplerEnv-WidowX基准上一致优于评估的VLA基线,在相同训练预算下达到最高的平均成功率。在LIBERO-Plus的OOD扰动下,LaMP展现出更强的鲁棒性,相比最强先验方法平均提升9.7%。项目页面见:https://lamppolicy.github.io。

英文摘要:

We introduce \textbf{LaMP}, a dual-expert Vision-Language-Action framework that embeds dense 3D scene flow as a latent motion prior for robotic manipulation.Existing VLA models regress actions directly from 2D semantic visual features, forcing them to learn complex 3D physical interactions implicitly.This implicit learning strategy degrades under unfamiliar spatial dynamics.LaMP addresses this limitation by aligning a flow-matching \emph{Motion Expert} with a policy-predicting \emph{Action Expert} through gated cross-attention.Specifically, the Motion Expert generates a one-step partially denoised 3D scene flow, and its hidden states condition the Action Expert without full multi-step reconstruction.We evaluate LaMP on the LIBERO, LIBERO-Plus, and SimplerEnv-WidowX simulation benchmarks as well as real-world experiments.LaMP consistently outperforms evaluated VLA baselines across LIBERO, LIBERO-Plus, and SimplerEnv-WidowX benchmarks, achieving the highest reported average success rates under the same training budgets. On LIBERO-Plus OOD perturbations, LaMP shows improved robustness with an average 9.7\% gain over the strongest prior baseline.Our project page is available at https://summerwxk.github.io/lamp-project-page/.

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。

↑