arXiv 2506.140092026-05-15cs.RO

GRaD-Nav++: 基于视觉-语言模型的视觉无人机导航：高斯辐射场与可微动力学

GRaD-Nav++: Vision-Language Model Enabled Visual Drone Navigation with Gaussian Radiance Fields and Differentiable Dynamics

Department of Mechanical Engineering, Stanford University（斯坦福大学机械工程系）；Department of Aeronautics and Astronautics, Stanford University（斯坦福大学航空航天工程系）

Qianzhong Chen, Naixiang Gao, Suning Huang, JunEn Low, Timothy Chen, Jiankai Sun, Mac Schwager

备注：: Published in: IEEE Robotics and Automation Letters ( Volume: 11, Issue: 2, February 2026)
发表信息：: Chen, Qianzhong, et al. "Grad-nav++: Vision-language model enabled visual drone navigation with gaussian radiance fields and differentiable dynamics." IEEE Robotics and Automation Letters 11.2 (2025): 1418-1425
DOI：: 10.1109/LRA.2025.3643290

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

GRaD-Nav++提出一种轻量级视觉-语言-动作框架，通过可微强化学习在3D高斯点云模拟器中训练，实现基于自然语言指令的实时无人机导航，展示在多任务和多环境下的高效导航能力。

摘要：

自主无人机在无结构环境中解释并执行高层语言指令仍是一个长期目标。然而，现有方法受限于对人工技能的依赖、参数调优的繁琐或计算密集型模型无法用于机载使用。我们引入GRaD-Nav++，一种轻量级的视觉-语言-动作（VLA）框架，能够在机载环境中实时运行并执行自然语言指令。我们的策略在逼真3D高斯点云（3DGS）模拟器中通过可微强化学习（DiffRL）训练，能够高效学习低层控制，从视觉和语言输入中学习。其核心是一个专家混合（MoE）动作头，能够自适应地路由计算以提高泛化能力并缓解遗忘。在多任务泛化实验中，GRaD-Nav++在训练任务中达到83%的成功率，在未见过的任务中达到75%。在真实硬件部署中，其在训练任务中的成功率为67%，在未见过的任务中为50%。在多环境适应实验中，GRaD-Nav++在多样化的模拟环境中平均成功率为81%，在多样的真实世界设置中为67%。这些结果为完全机载视觉-语言-动作（VLA）飞行建立了新的基准，并证明了紧凑、高效的模型可以在不依赖外部基础设施的情况下实现可靠的、语言引导的导航。

英文摘要：

Autonomous drones capable of interpreting and executing high-level language instructions in unstructured environments remain a long-standing goal. Yet existing approaches are constrained by their dependence on hand-crafted skills, extensive parameter tuning, or computationally intensive models unsuitable for onboard use. We introduce GRaD-Nav++, a lightweight Vision-Language-Action (VLA) framework that runs fully onboard and follows natural-language commands in real time. Our policy is trained in a photorealistic 3D Gaussian Splatting (3DGS) simulator via Differentiable Reinforcement Learning (DiffRL), enabling efficient learning of low-level control from visual and linguistic inputs. At its core is a Mixture-of-Experts (MoE) action head, which adaptively routes computation to improve generalization while mitigating forgetting. In multi-task generalization experiments, GRaD-Nav++ achieves a success rate of 83% on trained tasks and 75% on unseen tasks in simulation. When deployed on real hardware, it attains 67% success on trained tasks and 50% on unseen ones. In multi-environment adaptation experiments, GRaD-Nav++ achieves an average success rate of 81% across diverse simulated environments and 67% across varied real-world settings. These results establish a new benchmark for fully onboard Vision-Language-Action (VLA) flight and demonstrate that compact, efficient models can enable reliable, language-guided navigation without relying on external infrastructure.

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。