Search papers, labs, and topics across Lattice.
This paper introduces CONTACTMIMIC, a learning framework that enhances humanoid robot interaction with objects by integrating part-level binary contact commands with keypoint tracking. By employing contact-following rewards and a trajectory augmentation scheme, the framework effectively decouples contact behavior from keypoint geometry, allowing for precise physical interactions. Simulation results demonstrate that CONTACTMIMIC not only achieves superior contact controllability across various manipulation tasks but also outperforms traditional keypoint-only trackers in contact-relevant scenarios.
Robots can now achieve precise physical interactions with objects without relying on task-specific rewards, thanks to a novel approach that decouples contact from keypoint tracking.
Keypoint tracking alone is insufficient for object interaction tasks such as sitting on a chair, wiping a board, or pushing furniture, where the robot can reach the correct pose without making meaningful physical contact with the object. We present CONTACTMIMIC, a learning framework that tracks explicit partlevel binary contact commands alongside keypoint trajectories. CONTACTMIMIC is made possible through the use of contact-following rewards and a trajectory augmentation scheme aimed at breaking the correlations between keypoint trajectories and contact labels. The resulting policy successfully decouples contact behavior from keypoint geometry, and achieves precise physical contact as well as contact-controllability (produce or suppress contact during deployment as desired). Simulation experiments across 10 diverse human-object interaction motions confirm that CONTACTMIMIC exhibits contact controllability that enables it to complete manipulation tasks without task-specific rewards, while also outperforming keypoint-only trackers on contact-relevant tasks. Ablations confirm the necessity of the proposed trajectory augmentation scheme and sim2real deployment validates contact controllability in the real world across 5 different motions. Video results are available on https://lixinyao11.github.io/contactmimic-page/.
