Hold My Beer: Learning Gentle Humanoid Locomotion and End-Effector Stabilization Control

TL;DR

This paper introduces SoFTA, a dual-frequency control framework for humanoids that improves end-effector stabilization during walking, enabling delicate tasks like carrying liquids without spilling.

Contribution

We propose a novel Slow-Fast Two-Agent framework that decouples upper and lower body control, enhancing stability and coordination in humanoid locomotion.

Findings

Reduces end-effector acceleration by 2-5x compared to baselines.

Achieves near human-level stability in delicate tasks.

Enables humanoids to carry full cups and capture steady video during walking.

Abstract

Can your humanoid walk up and hand you a full cup of beer, without spilling a drop? While humanoids are increasingly featured in flashy demos like dancing, delivering packages, traversing rough terrain, fine-grained control during locomotion remains a significant challenge. In particular, stabilizing a filled end-effector (EE) while walking is far from solved, due to a fundamental mismatch in task dynamics: locomotion demands slow-timescale, robust control, whereas EE stabilization requires rapid, high-precision corrections. To address this, we propose SoFTA, a Slow-Fast Two-Agent framework that decouples upper-body and lower-body control into separate agents operating at different frequencies and with distinct rewards. This temporal and objective separation mitigates policy interference and enables coordinated whole-body behavior. SoFTA executes upper-body actions at 100 Hz for precise EE control and lower-body actions at 50 Hz for robust gait. It reduces EE acceleration by 2-5x relative to baselines and performs much closer to human-level stability, enabling delicate tasks such as carrying nearly full cups, capturing steady video during locomotion, and disturbance rejection with EE stability.