外观
论文arXiv 2609.145431 个月前
NavPatch：使用 Vision-Language Models 的证据引导对象级代价地图校正
原文链接：NavPatch: Evidence-Guided Object-Level Costmap Correction with Vision-Language Models（arXiv 2609.14543 · 发布日期：2026-09-13）
NavPatch: Evidence-Guided Object-Level Costmap Correction with Vision-Language Models
针对几何地图障碍表示与导航需求不匹配，NavPatch 用 VLM 进行对象级代价地图校正，并在五种布局的真机试验中验证。
核心要点
- 几何地图可能漏掉低矮电缆、误将柔性帘子视为阻挡，或低估交通锥需避让范围。
- NavPatch 周期性用 VLM 理解场景，为导航相关物体类别分配 ADD、REMOVE 或 EXTEND。
- 开放词汇定位物体实例，LiDAR 与 RGB-D 提供 3D 支持，结合质量过滤和跨帧维护管理补丁。
- 50 次真机试验总体成功率 86.0%；200 次消融显示成功率由 70.0% 提至 86.0%，误提交率降低。
验证范围与局限：验证范围为五种布局中的 50 次真机试验，以及四种配置共 200 次消融运行。
英文摘要 (Abstract)
Mobile robots typically rely on geometric maps for obstacle avoidance and path planning, but the resulting obstacle representation does not always match how an object should affect navigation. A low lying cable may be missed, a flexible curtain may create spurious blockage, and a traffic cone may require an exclusion region larger than its observed footprint. We present NavPatch, an object level correction layer that assigns ADD, REMOVE, or EXTEND to navigation relevant object categories through periodic scene understanding with a vision-language model. Open vocabulary grounding localizes object instances, and LiDAR and RGB-D observations provide 3D support. Observation quality filtering and cross frame maintenance determine when each correction patch is committed, replaced, or revoked. In 50 real robot trials across five layouts, NavPatch achieves an overall success rate of 86.0%. An ablation study of four configurations with 200 runs in total shows that NavPatch improves the success rate from 70.0% to 86.0% and reduces the false commit rate from 68.4% to 40.7% compared with updates based only on the current observation.
作者：Shiji Sun, Xingyu Tao, Hao Wang, Ling Wang, Zhengyi Chen
来源
本页事实对应的公开页面，以原文为准。