MOSU：基于多模态场景理解的自主长距离机器人导航
MOSU: Autonomous Long-range Robot Navigation with Multi-modal Scene Understanding
AI总结：
MOSU系统通过融合LiDAR几何、图像语义和视觉语言模型实现多模态场景理解，结合GPS与QGIS全局规划，在GND基准上提升10%可通行性并保持相当导航距离。
摘要：
我们提出MOSU，一种新颖的自主长距离导航系统，通过多模态感知和道路场景理解增强移动机器人的全局导航能力。MOSU通过整合几何、语义和上下文信息来应对室外机器人导航挑战，以确保全面的场景理解。该系统结合GPS和基于QGIS地图的路径规划进行高层全局路径规划，并利用多模态轨迹生成进行局部导航细化。在轨迹生成方面，MOSU利用多种模态：基于LiDAR的几何数据用于精确避障，基于图像的语义分割用于可通行性评估，以及视觉语言模型（VLMs）来捕捉社会上下文，使机器人能够在复杂环境中遵守社会规范。这种多模态集成改善了场景理解并增强了可通行性，使机器人能够适应多样的室外条件。我们在真实世界道路环境中评估了我们的系统，并在GND数据集上进行了基准测试，在可导航地形上实现了10%的可通行性提升，同时保持了与现有全局导航方法相当的导航距离。
英文摘要：
We present MOSU, a novel autonomous long-range navigation system that enhances global navigation for mobile robots through multimodal perception and on-road scene understanding. MOSU addresses the outdoor robot navigation challenge by integrating geometric, semantic, and contextual information to ensure comprehensive scene understanding. The system combines GPS and QGIS map-based routing for high-level global path planning and multi-modal trajectory generation for local navigation refinement. For trajectory generation, MOSU leverages multi-modalities: LiDAR-based geometric data for precise obstacle avoidance, image-based semantic segmentation for traversability assessment, and Vision-Language Models (VLMs) to capture social context and enable the robot to adhere to social norms in complex environments. This multi-modal integration improves scene understanding and enhances traversability, allowing the robot to adapt to diverse outdoor conditions. We evaluate our system in real-world on-road environments and benchmark it on the GND dataset, achieving a 10% improvement in traversability on navigable terrains while maintaining a comparable navigation distance to existing global navigation methods.
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。
