EgoPush：面向移动机器人的端到端第一人称多物体重排学习
EgoPush: Egocentric Multi-Object Rearrangement for Mobile Robots via Constrained Teacher Observability
AI总结：
EgoPush通过端到端学习实现移动机器人第一人称多物体重排，利用感知驱动的潜在空间和阶段分解奖励提升重排效率与现实迁移能力。
摘要：
人类能够利用第一人称感知在杂乱环境中重新排列物体，无需全局坐标即可导航遮挡。受此能力启发，我们研究了使用单个第一人称摄像头进行长视距多物体非抓取重排的移动机器人问题。我们引入了EgoPush，一种策略学习框架，能够实现第一人称、感知驱动的重排，而无需依赖通常在动态场景中失效的显式全局状态估计。EgoPush设计了一个以物体为中心的潜在空间，用于编码物体之间的相对空间关系，而不是绝对姿态。这种设计使一个特权强化学习（RL）教师能够从稀疏关键点中联合学习潜在状态和移动动作，然后将其蒸馏成纯视觉的学生策略。为减少教师与部分观察学生之间的监督差距，我们限制教师的观察到视觉可访问的线索。这诱导出可从学生视角恢复的主动感知行为。为解决长视距信用分配问题，我们通过时间衰减的、阶段局部完成奖励将重排分解为阶段级子问题。大量模拟实验表明，EgoPush在成功率上显著优于端到端RL基线，消融研究验证了每个设计选择。我们进一步在真实世界移动平台上展示了零样本仿真到现实的迁移。代码和视频可在https://ai4ce.github.io/EgoPush/上获取。
英文摘要：
Humans rearrange objects in cluttered environments using egocentric perception, actively moving to keep task-relevant spatial cues in view. Mobile robots have not matched this: rearrangement is usually built on a global pose estimate or a map, which is exactly what a robot carrying one camera lacks, while pushing keeps changing the scene it would have to be built from. We present EgoPush, which pushes objects into anchor-relative formations from onboard RGB-D alone, with no global localization, external tracking, or map at deployment, and transfers zero-shot to a TurtleBot in controlled and visually cluttered scenes. What makes this learnable turns out to be a property of the teacher rather than of the student: three privileged teachers trained with identical rewards, architecture, and hyperparameters all exceed $98\%$ success, yet their distilled egocentric students reach $0\%$, $54.8\%$, and $87.3\%$, the only variable being the teacher's observation function. EgoPush therefore trains the teacher under egocentric observability constraints, restricting it to visibility-limited cues and revealing target references only when the anchor is centrally visible, so that its supervision is recoverable by a depth-based student distilled online. Making the teacher trainable in the first place needs two further pieces: a role-grouped object-centric interface shared by teacher and student, and stage-wise temporally decayed rewards for long-horizon credit assignment. Videos, the playable task, and code are available at https://ai4ce.github.io/EgoPush/.
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。
