RoTri-Diff: A Spatial Robot–Object Triadic Interaction-Guided Diffusion Model for Bimanual Manipulation
Zixuan Chen, Nga Teng Chan, Yiwen Hou, Chenrui Tie, Zixuan Liu, Haonan Chen, Junting Chen, Jieqi Shi, Yang Gao, Jing Huo, Lin Shao · 2026
Read RoTri-Diff: A Spatial Robot–Object Triadic Interaction-Guided Diffusion Model for Bimanual Manipulation by Zixuan Chen, Nga Teng Chan, Yiwen Hou, Chenrui Tie, Zixuan Liu, Haonan Chen, Junting Chen, Jieqi Shi, Yang Gao, Jing Huo, Lin Shao (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2603-07165.txt.
Description
Chen and colleagues improve two-arm robot manipulation with RoTri-Diff, a diffusion model guided by robot-object triadic interaction. Past imitation methods, whether robot-centered or object-centered, miss the changing geometric ties among the two arms and the object. That gap causes collisions, unstable grasps, and weak performance on hard tasks. RoTri encodes the relative 6D poses of both arms and the object as triangular constraints. RoTri-Diff combines those constraints with keyposes and object motion in a layered diffusion process.
Catalog record
| CREATOR | Zixuan Chen, Nga Teng Chan, Yiwen Hou, Chenrui Tie, Zixuan Liu, Haonan Chen, Junting Chen, Jieqi Shi, Yang Gao, Jing Huo, Lin Shao |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 36002 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2603.07165