CMP: Robust Whole-Body Tracking for Loco-Manipulation via Competence Manifold Projection

TL;DR

This paper introduces CMP, a novel approach for robust whole-body control in legged manipulators, significantly improving safety and performance under out-of-distribution inputs through competence manifold projection.

Contribution

It proposes a competence manifold projection method with a safety scheme and latent space alignment to enhance robustness against OOD perturbations.

Findings

Up to 10-fold increase in survival rate under OOD scenarios.

Less than 10% tracking degradation in challenging conditions.

Emergent generalization behaviors for OOD goal achievement.

Abstract

While decoupled control schemes for legged mobile manipulators have shown robustness, learning holistic whole-body control policies for tracking global end-effector poses remains fragile against Out-of-Distribution (OOD) inputs induced by sensor noise or infeasible user commands. To improve robustness against these perturbations without sacrificing task performance and continuity, we propose Competence Manifold Projection (CMP). Specifically, we utilize a Frame-Wise Safety Scheme that transforms the infinite-horizon safety constraint into a computationally efficient single-step manifold inclusion. To instantiate this competence manifold, we employ a Lower-Bounded Safety Estimator that distinguishes unmastered intentions from the training distribution. We then introduce an Isomorphic Latent Space (ILS) that aligns manifold geometry with safety probability, enabling efficient O(1) seamless defense against arbitrary OOD intents. Experiments demonstrate that CMP achieves up to a 10-fold survival rate improvement in typical OOD scenarios where baselines suffer catastrophic failure, incurring under 10% tracking degradation. Notably, the system exhibits emergent ``best-effort'' generalization behaviors to progressively accomplish OOD goals by adhering to the competence boundaries. Result videos are available at: https://shepherd1226.github.io/CMP.