HAF: Adapting Generalist VLAs to Humanoid Whole-Body Loco-manipulation via Hierarchical Action Flow and Spectral Latent RL
2026-08-17 • Robotics
RoboticsArtificial Intelligence
AI summaryⓘ
The authors developed HAF, a new system to help humanoid robots better coordinate movements like walking and using their arms together. They found that existing vision-language-action models struggle with the complexity of whole-body motions when controlled all at once. HAF breaks down the full-body action into stages and uses a special learning method to improve robot performance safely and efficiently without retraining large parts of the model. Tested on real robots, their approach worked better than simpler methods for various tasks involving movement and manipulation.
Humanoid robotsVision-language-action (VLA) modelsWhole-body loco-manipulationBehavior cloningReinforcement learningFlow matchingHierarchical action generationSoft actor-critic (SAC)Dimensionality reductionOffline-to-online learning
Authors
Langzhe Gu, Chengkai Hou, Meng Li, Xinhua Wang, Jiaming Liu, Xinyuan Lv, Bowei Zhang, Shuanghao Bai, Guangrun Li, Jingyang He, Gaole Dai, Ziluo Ding, Zhiyuan Xu, Kuan Cheng, Jian Tang, Zhengping Che, Shanghang Zhang
Abstract
Humanoid robots hold great promise as general-purpose agents in human-centered environments, yet generalist vision-language-action (VLA) foundation models are not readily applicable to humanoid whole-body loco-manipulation. The high dimensionality and interdependence of humanoid motions make it challenging for conventional single-stage VLA architectures to coordinate locomotion, waist posture, and dual-arm manipulation effectively. Moreover, policies trained through offline behavior cloning can remain suboptimal during real-world deployment. Although online reinforcement learning can refine policies through real-world interaction, directly tuning large VLA backbones demands excessive computation and may introduce safety risks during real-robot exploration. To address these bottlenecks, we introduce HAF (Humanoid Adaptation Framework), a two-part framework consisting of HAF-VLA and HAF-Steer that transfers off-the-shelf generalist VLA foundation models to humanoid whole-body loco-manipulation. HAF-VLA is a hierarchical action-flow generator built on a pretrained flow-matching VLA. It splits full-body action denoising into three sequential stages with stage embeddings and cross-stage KV caches that retain kinematic dependencies, avoiding incoherent whole-body actions from one-shot generation. On top of the frozen HAF-VLA, HAF-Steer is a latent offline-to-online RL pipeline that leverages flow-matching invertibility and DCT-based dimensionality reduction to restrict RL optimization to a compact noise subspace and train a regularized SAC policy. This avoids updating the large VLA backbone and enables efficient real-world policy refinement. Evaluated on seven real-world humanoid loco-manipulation tasks, HAF surpasses vanilla single-stage VLA baselines and improves whole-body coordination and task performance. Project website: https://grange007.github.io/HAF .