Before the Body Moves: Learning Anticipatory Joint Intent for Language-Conditioned Humanoid Control

TL;DR

This paper introduces DAJI, a hierarchical framework enabling humanoid robots to anticipate future physical states from language instructions, improving control and responsiveness.

Contribution

DAJI is the first framework to explicitly encode anticipatory joint intent from language, integrating future-aware control with language-conditioned humanoid motion.

Findings

Achieves 94.42% success in HumanML3D-style generation

Attains 0.152 subsequence FID on BABEL dataset

Demonstrates strong anticipatory control in streaming instructions

Abstract

Natural language is an intuitive interface for humanoid robots, yet streaming whole-body control requires control representations that are executable now and anticipatory of future physical transitions. Existing language-conditioned humanoid systems typically generate kinematic references that a low-level tracker must repair reactively, or use latent/action policies whose outputs do not explicitly encode upcoming contact changes, support transfers, and balance preparation. We propose \textbf{DAJI} (\emph{Dynamics-Aligned Joint Intent}), a hierarchical framework that learns an anticipatory joint-intent interface between language generation and closed-loop control. DAJI-Act distills a future-aware teacher into a deployable diffusion action policy through student-driven rollouts, while DAJI-Flow autoregressively generates future intent chunks from language and intent history. Experiments show that DAJI achieves strong results in anticipatory latent learning, single-instruction generation, and streaming instruction following, reaching 94.42\% rollout success on HumanML3D-style generation and 0.152 subsequence FID on BABEL.