ULTRA: Unified Multimodal Control for Autonomous Humanoid Whole-Body Loco-Manipulation

TL;DR

ULTRA is a unified framework enabling humanoid robots to perform autonomous, goal-directed whole-body loco-manipulation using perception and high-level commands, overcoming previous limitations in data quality, skill scalability, and reliance on motion references.

Contribution

The paper introduces ULTRA, combining physics-based neural retargeting and a multimodal controller to generate behavior from perception and high-level tasks, advancing humanoid loco-manipulation.

Findings

ULTRA outperforms tracking-only baselines in simulation and real-world tests.

It generalizes to goal-conditioned tasks from egocentric perception.

ULTRA maintains robustness under out-of-distribution scenarios.

Abstract

Achieving autonomous and versatile whole-body loco-manipulation remains a central barrier to making humanoids practically useful. Yet existing approaches are fundamentally constrained: retargeted data are often scarce or low-quality; methods struggle to scale to large skill repertoires; and, most importantly, they rely on tracking predefined motion references rather than generating behavior from perception and high-level task specifications. To address these limitations, we propose ULTRA, a unified framework with two key components. First, we introduce a physics-driven neural retargeting algorithm that translates large-scale motion capture to humanoid embodiments while preserving physical plausibility for contact-rich interactions. Second, we learn a unified multimodal controller that supports both dense references and sparse task specifications, under sensing ranging from accurate motion-capture state to noisy egocentric visual inputs. We distill a universal tracking policy into this controller, compress motor skills into a compact latent space, and apply reinforcement learning finetuning to expand coverage and improve robustness under out-of-distribution scenarios. This enables coordinated whole-body behavior from sparse intent without test-time reference motions. We evaluate ULTRA in simulation and on a real Unitree G1 humanoid. Results show that ULTRA generalizes to autonomous, goal-conditioned whole-body loco-manipulation from egocentric perception, consistently outperforming tracking-only baselines with limited skills.