PhysVLMM：机器人场景图中铰接对象的交互条件物理信念图
PhysVLMM: Interaction-Conditioned Physical Belief Maps for Articulated Objects in Robot Scene Maps
Dibyendu Ghosh
中文摘要
移动机器人必须开门和抽屉，但现有场景图只记录物体是什么、是否移动过，不记录它对拉拽会如何响应。PhysVLMM 为视觉-语言-运动图（Vision-Language-Motion Map）中每个铰接部件维护物理信念，并回答规划器查询，包含地图更新与查询两个算法。观测部件运动无法辨识其动力学，一次测量拉拽只能辨识被拉部件、且方向至多确定一个。因此该方法（i）将该方向保留为后验而非点估计，（ii）以闭式形式把每次拉拽的证据传递给地图中未接触的相似部件，（iii）在给定风险下回答动作是否成功，或弃权。在 18 个家庭的真实门和抽屉上，留出拉拽中关节类型识别率 86%（学习式单视图方法为 48%），拉拽响应预测误差比可微仿真低 41%（CRPS 0.323 对 0.544），速度快四个数量级（每例 13 ms 对 230 s）。后验将最终位置误差从 0.460 降至 0.270，证据共享将未接触部件的力误差从 56% 降至 42%，在 0.02 至 0.30 的每个给定风险下错误安全回答均未超出该风险。汇总一个家庭所有已测部件与相似部件分组效果相当，更简单变体预测被拉部件停止位置至少同样好。
关键要点
- 01问题：机器人场景图只记录物体身份与是否移动，缺少对拉拽响应的物理预测能力。
- 02方法：为每个铰接部件维护方向后验，以闭式形式把拉拽证据共享给相似部件，并按风险回答或弃权。
- 03结果：18 个家庭真实门抽屉上关节类型识别率 86%，响应预测误差比可微仿真低 41%，速度快四个数量级。
- 04结果：后验将最终位置误差从 0.460 降至 0.270，证据共享将未接触部件力误差从 56% 降至 42%。
- 05局限：单次拉拽只能辨识被拉部件且方向至多确定一个，观测运动无法辨识动力学。
解读
尚无解读。
原始英文摘要
arXiv:2610.40539v1 Announce Type: new Abstract: Mobile robots must open doors and drawers, yet their scene maps record what each object is and whether it has moved, not how it will respond to a pull. This paper proposes PhysVLMM, a method that maintains a physical belief for every articulated part of a Vision-Language-Motion Map and answers a planner's queries about it, with two algorithms: a map update and a query. Watching a part move provably cannot identify its dynamics, and a measured pull identifies only the pulled part, up to one direction. PhysVLMM therefore (i) keeps that direction as a posterior instead of a point estimate, (ii) passes the evidence of each pull in closed form to the untouched look-alike parts of the map, and (iii) answers whether an action will succeed at a stated risk, or abstains. On real doors and drawers in 18 homes, it identifies the joint type in 86% of held-out pulls (48% for a learned single-view method) and predicts the response to a pull with 41% lower error than differentiable simulation (CRPS 0.323 vs. 0.544), four orders of magnitude faster (13 ms vs. 230 s per case). The posterior lowers the final-position error from 0.460 to 0.270, sharing lowers the force error of untouched parts from 56% to 42%, and at every stated risk from 0.02 to 0.30 the false-safe answers stay within it. Pooling all measured parts of a home does as well as the look-alike groups, and simpler variants predict where a pulled part stops at least as well.