InfCoiL: Coordinated Planner-Controller Learning for Closed-Loop Physics-Based Human-Object Interaction
Make generated motion executable. InfCoiL lets a flow-matching planner and a physics-based controller improve each other through shared rollout feedback, alternating their RL updates when progress plateaus.
Watch demos
Closed-loop table transport
Replanning through tripod transport
BibTeX
@inproceedings{zou2026infcoil,
title = {{InfCoiL}: Coordinated Planner-Controller Learning for Closed-Loop Physics-Based Human-Object Interaction},
author = {Zou, Yude and Yao, Yifei and Hao, Yiwei and Wang, Hanqing and Gao, Xing},
booktitle = {Advances in Neural Information Processing Systems},
year = {2026}
}