CoRMA: Contrastive RMA for Contact-Rich Meta-Adaptation

TL;DR

CoRMA introduces a contact-based meta-adaptation framework that infers semantic contact context online, enabling effective within-episode adaptation for force-dominant robotic assembly tasks without requiring demonstrations or gradient updates.

Contribution

It replaces raw parameter adaptation with a semantic contact context inferred by a Transformer, improving real-world success in assembly tasks under noise.

Findings

CoRMA outperforms FORGE baselines on hardware with higher real success.

Semantic contact inference enables effective within-episode adaptation.

The framework generalizes across multiple assembly tasks.

Abstract

We present CoRMA(Contrastive Robotic Motor Adaptation), a context-based meta-adaptation framework that modifies RMA for force-dominant assembly. CoRMA replaces raw simulator-parameter adaptation with a compact 6D simulator-only semantic contact context describing contact onset, lateral engagement, guided transition, contact direction, and jamming. A deployable causal Transformer adapter infers this context online from force, proprioceptive, and action histories using semantic regression and a force-regime contrastive objective. At deployment, oracle context is removed and replaced by the inferred context, enabling within-episode adaptation without demonstrations, privileged inputs, or gradient updates. We evaluate CoRMA on PegInsert, GearMesh, and NutThread in Isaac Lab / Isaac Sim~5.0 and on a real Marvin arm. Compared with FORGE baselines that achieve high simulation success but degrade substantially on hardware, CoRMA retains higher verified real success under controlled target-pose noise. These results support semantic contact inference as a reusable adaptation interface within a related assembly task family, while broader unseen-task generalization and Real2Sim calibration remain future work.