Precise SE(3) End-Effector Tracking in Whole-Body Humanoid Control

2026-10-07

AI Summary

This study addresses the challenge of precise end-effector tracking in humanoid whole-body control, which is severely hindered by floating-base oscillations, gravity, and dynamic coupling. To overcome these limitations, this work proposes ResGAC, a controller that integrates geometric admittance control with residual reinforcement learning to compensate for unmodeled dynamics. The framework unifies manipulation reference frames through a left-invariant geometric formulation and introduces a ground-attached heading frame to eliminate pelvic pose interference. Furthermore, it achieves coordinated motion and balance via SE(3) task-space feedback combined with a shared joint action space. Experimental evaluations on the Unitree G1 platform demonstrate that ResGAC attains a 90% success rate in peg-in-hole tasks, significantly outperforming the SONIC baseline at 50%, thereby validating its capability for high-precision pose tracking.

Abstract

Precise end-effector tracking during humanoid whole-body motion is challenging due to floating-base oscillations, gravity, dynamic coupling, and locomotion-induced disturbances. We propose ResGAC, a whole-body humanoid controller for precise end-effector pose tracking that combines geometric admittance control (GAC) with residual reinforcement learning. GAC provides structured $\SE$ task-space feedback and generates nominal arm joint-position targets, while residual RL compensates for unmodeled dynamics and coordinates locomotion and balance in the shared joint-position action space. The left-invariant geometric formulation allows the same GAC law to be used across manipulation reference frames. This enables the use of a ground-attached heading frame that preserves planar locomotion while removing pelvis roll, pitch, and heave from the manipulation reference, thereby reducing reference-induced end-effector motion during locomotion. ResGAC is validated on a real Unitree G1 humanoid. Across four standing end-effector tracking benchmarks, ResGAC consistently outperforms representative baselines, including SONIC, achieving lower translational and rotational errors. Real-world experiments further demonstrate reduced propagation of pelvis motion to the desired end-effector pose using the proposed ground-attached heading frame. ResGAC achieves $90\%$ success in a standing peg-in-hole task compared with $50\%$ for SONIC, and accurate world-frame $\SE$ end-effector pose tracking during lower-body motion. Experimental videos are included in the supplementary material and are also available on the project website: https://resgac.github.io/ResGAC-website/.

Problem

• Humanoid whole-body control
• End-effector tracking
• SE(3) pose tracking
• Floating-base oscillations
• Dynamic coupling

Innovation

• SE(3) end-effector tracking
• geometric admittance control
• residual reinforcement learning
• whole-body humanoid control
• left-invariant geometry