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 2603.08862cs.ROcs.LG
收藏

APPLV: 从视觉-语言-动作模型中自适应学习规划器参数

APPLV: Adaptive Planner Parameter Learning from Vision-Language-Action Model

Department of Computer Science, George Mason University(计算机科学系,乔治·马歇尔大学);Department of Engineering Science, University of South Florida(工程科学系,佛罗里达州立大学);Department of Computer Science, Rutgers University(计算机科学系,罗格斯大学)

Yuanjie Lu, Beichen Wang, Zhengqi Wu, Yang Li, Xiaomin Lin, Chengzhi Mao, Xuesu Xiao

AI总结:

本文提出APPLV,通过从视觉-语言-动作模型中自适应学习规划器参数,提升移动机器人在受限环境中的导航性能与泛化能力。

摘要:

在高度受限的环境中实现自主导航仍然对移动机器人来说是一个挑战。经典导航方法提供安全保证,但需要环境特定的参数调节;端到端学习避免了参数调节,但在受限空间中难以实现精确控制。为此,最近的机器人学习方法自动化了参数调节,同时保留了经典系统的安全性,但仍面临在未见环境中泛化的能力挑战。最近,视觉-语言-动作(VLA)模型通过利用基础模型的场景理解能力显示出潜力,但在导航任务中仍然面临精确控制和推理延迟的挑战。在本文中,我们提出了从视觉-语言-动作模型中自适应学习规划器参数(APPLV)。与传统VLA模型直接输出动作不同,APPLV利用预训练的视觉-语言模型,结合回归头来预测配置经典规划器的参数。我们开发了两种训练策略:监督学习微调从收集的导航轨迹中学习,以及强化学习微调以进一步优化导航性能。我们评估了APPLV在多个运动规划器上的性能,并在模拟的Benchmark Autonomous Robot Navigation(BARN)数据集以及物理机器人实验中进行了测试。结果表明,APPLV在导航性能和对未见环境的泛化能力上均优于现有方法。

英文摘要:

Autonomous navigation in highly constrained environments remains challenging for mobile robots. Classical navigation approaches offer safety assurances but require environment-specific parameter tuning; end-to-end learning bypasses parameter tuning but struggles with precise control in constrained spaces. To this end, recent robot learning approaches automate parameter tuning while retaining classical systems' safety, yet still face challenges in generalizing to unseen environments. Recently, Vision-Language-Action (VLA) models have shown promise by leveraging foundation models' scene understanding capabilities, but still struggle with precise control and inference latency in navigation tasks. In this paper, we propose Adaptive Planner Parameter Learning from Vision-Language-Action Model (\textsc{applv}). Unlike traditional VLA models that directly output actions, \textsc{applv} leverages pre-trained vision-language models with a regression head to predict planner parameters that configure classical planners. We develop two training strategies: supervised learning fine-tuning from collected navigation trajectories and reinforcement learning fine-tuning to further optimize navigation performance. We evaluate \textsc{applv} across multiple motion planners on the simulated Benchmark Autonomous Robot Navigation (BARN) dataset and in physical robot experiments. Results demonstrate that \textsc{applv} outperforms existing methods in both navigation performance and generalization to unseen environments.

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。

↑