A Hierarchical Framework for Humanoid Locomotion with Supernumerary Limbs
Bowen Zhi
2025-12-02
Summary
This research focuses on making humanoid robots more stable when they have extra limbs, often called Supernumerary Limbs (SLs). Adding these limbs throws off the robot's balance, and this project aims to fix that.
What's the problem?
When you attach extra limbs to a humanoid robot, it becomes much harder to keep it from falling over. The movement and weight of these extra limbs create disturbances that disrupt the robot’s natural walking motion and balance. Essentially, the robot struggles to adjust to the added dynamics.
What's the solution?
The researchers developed a control system with two main parts working together. First, they used imitation learning and curriculum learning to teach the robot a basic walking gait. Then, they added a second layer that actively uses the extra limbs to help the robot regain its balance as it walks. This system separates the walking motion from the balancing adjustments, allowing each to be optimized independently.
Why it matters?
This work is important because it allows for the creation of more capable humanoid robots. By solving the stability problem with extra limbs, robots can potentially carry more, manipulate objects in more complex ways, and adapt to uneven terrain more effectively. It’s a step towards robots that can perform more useful tasks in the real world.
Abstract
The integration of Supernumerary Limbs (SLs) on humanoid robots poses a significant stability challenge due to the dynamic perturbations they introduce. This thesis addresses this issue by designing a novel hierarchical control architecture to improve humanoid locomotion stability with SLs. The core of this framework is a decoupled strategy that combines learning-based locomotion with model-based balancing. The low-level component consists of a walking gait for a Unitree H1 humanoid through imitation learning and curriculum learning. The high-level component actively utilizes the SLs for dynamic balancing. The effectiveness of the system is evaluated in a physics-based simulation under three conditions: baseline gait for an unladen humanoid (baseline walking), walking with a static SL payload (static payload), and walking with the active dynamic balancing controller (dynamic balancing). Our evaluation shows that the dynamic balancing controller improves stability. Compared to the static payload condition, the balancing strategy yields a gait pattern closer to the baseline and decreases the Dynamic Time Warping (DTW) distance of the CoM trajectory by 47\%. The balancing controller also improves the re-stabilization within gait cycles and achieves a more coordinated anti-phase pattern of Ground Reaction Forces (GRF). The results demonstrate that a decoupled, hierarchical design can effectively mitigate the internal dynamic disturbances arising from the mass and movement of the SLs, enabling stable locomotion for humanoids equipped with functional limbs. Code and videos are available here: https://github.com/heyzbw/HuSLs.