Humanoid SRL: Augmenting Humanoids with Supernumerary Limbs for Coordinated Loco-Manipulation
In this study, we develop and integrate a supernumerary arm into a humanoid platform to investigate how morphological augmentation reshapes its feasible action space.
ProofPaper ↗
Key points
- Humanoid robots are envisioned to perform diverse human-level tasks through their anthropomorphic morphology.
- Supernumerary robotic limbs (SRLs) offer a way to extend these capabilities, but their integration requires the added limb to be coordinated with whole-body motion and balance.
- Through quantitative analysis and hardware experiments across two configurations, we show how the additional limb expands whole-body interaction capabilities by enabling manipulation and support to be coordinated beyond the constraints of the native dual-arm configuration.
- To control this augmented morphology, we leverage model-based trajectory optimization (TO) to generate dynamically feasible motions for goal-conditioned reinforcement learning (RL) controllers.
Sources (1)
- [1]Humanoid SRL: Augmenting Humanoids with Supernumerary Limbs for Coordinated Loco-ManipulationarXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 19, 01:08 AM
In this study, we develop and integrate a supernumerary arm into a humanoid platform to investigate how morphological augmentation reshapes its feasible action space.
Humanoid robots are envisioned to perform diverse human-level tasks through their anthropomorphic morphology.
Extractive summary: sentences quoted from the sources.
Before this
- Oct 19, 2026HarnessRL: Co-Evolving Policies with a Just-in-Time Training Harness for Open-Ended RL
- Oct 17, 2026Touch Where It Matters: Learning Local Tactile Policies for Fine-Grained Gripper Control
- Oct 17, 2026Gen2Dex: Learning Dexterous Manipulation from Simulation-Initialized Generated Video Demonstrations
- Oct 16, 2026Blind Precise Assembly: Proprioceptive Sim-to-Real Peg-in-Hole for Diverse Parts
- Oct 16, 2026Learning Robot Metareasoning: Adaptive Computation Budgets for Real-Time Control
- Oct 16, 2026Ego2Contact: Predicting Touch from Egocentric Human Video for Contact-Rich Robot Manipulation