arXiv cs.AI· Jiawei Geng, Zhenyang Liu, Yuzheng Wu, Yanwei Fu·· 12 小时前AI 评分31
AWAM：面向鲁棒机器人操作的活动轨迹条件视角想象
AWAM: Active Trajectory-Conditioned View Imagination for Robust Robot Manipulation
AI 导读
AWAM 是一种主动世界-动作模型，先预测任务相关的虚拟相机运动，将其作为结构化查询输入几何条件世界模型，生成可复用的空间先验并压缩为固定记忆，再通过 cross-attention 与实时观测融合，用于扩散式动作预测。在 LIBERO-Occ 上成功率达 75.9%，LIBERO-Plus 为 62.4%，RLBench-OG Occlusion1 为 76.5%，真实机器人试验为 94.4%。
正文 · 原文
Abstract:Rapid progress in embodied intelligence has enabled increasingly capable robots to follow language instructions and manipulate diverse objects. However, strong results in structured environments do not automatically transfer to open-world conditions, where clutter, occlusion, and changing viewpoints can obscure the spatial relationships required for action. To address this perceptual gap, a virtual active-perception paradigm is introduced. AWAM, an Active World-Action Model actively plans where to observe and imagines what will be seen along the planned trajectory before acting. Task-related virtual camera motion is predicted and used as a structured query into a geometry-conditioned world model, producing a reusable spatial prior for manipulation, which is later compressed into a fixed memory and fused with live observations through cross-attention for diffusion-based action prediction. AWAM achieves competitive success rates with 75.9% on LIBERO-Occ, 62.4% on LIBERO-Plus, 76.5% on RLBench-OG Occlusion1, and 94.4% in real-robot trials. These results support improved robustness to the evaluated occlusions and distribution shifts, addressing part of the gap between structured manipulation and less controlled environments.
| Subjects: | Robotics (cs.RO); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.23396 [cs.RO] |
| (or arXiv:2610.23396v1 [cs.RO] for this version) | |
| https://doi.org/10.48550/arXiv.2610.23396 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Jiawei Geng [view email]
[v1]
Fri, 16 Oct 2026 12:53:08 UTC (14,345 KB)
来源：arXiv cs.AI · arxiv.org