AI

CoDance: Learning Reactive and Compliant Human-Humanoid Interaction from Video

Read full article at source ↗

Based on a feed excerpt; full article content has not been acquired. Source evidence retrieved 08/10/2026, 15:18:24 UTC.

Briefings

CoDance is a framework for learning reactive and compliant human-humanoid interaction from video, using partnered dancing as a challenging instantiation. The system retargets motions into a robot reference and a moving partner, introducing a multi-link compliance augmentation that transforms kinematic demonstrations into force-aware training data. Policies trained on this data follow the observed partner while preserving locomotion style and responding to physical interaction. In simulation, policies reproduce approximately 80% of wrist displacement; on physical humanoid, CoDance enables sustained two-hand dancing with repeated forward and backward transitions.

Key points

• CoDance framework uses partnered dancing as a challenging instantiation for learning reactive and compliant human-humanoid interaction from video.
• Multi-link compliance augmentation transforms kinematic demonstrations into force-aware training data, enabling policies to follow partners while preserving locomotion style.

Context

This research presents a novel framework for learning human-humanoid interaction from video, utilizing partnered dancing as a challenging instantiation. The study introduces a multi-link compliance augmentation that adapts robot references under structured forces, allowing policies to maintain stable and natural movement while responding to physical interaction forces.

Limitations

Simulation results show policies reproduce approximately 80% of wrist displacement. Physical testing on a humanoid robot enabled sustained two-hand dancing with repeated transitions between forward and backward motions, but specific physical metrics or failure modes beyond this are not reported.

Generated by Bittrees-hosted qwen35-2b sha256:aaf42c8b7c3c on 09/11/2026, 03:38:32 UTC. Based on publisher-provided feed evidence; may not cover the full original article. Check the source for details.

View archived briefing on IPFS
bafybeifxque3rqity6fbgaqaowxcyp6ztyyp6kkxnhcydko75srbfvn3bq