Active Embodiment Identification with Reinforcement Learning for Legged Robots

2026/05/08 by Nico Bohlinger, Jan Peters · 1 voice
Computer Science · Engineering · #Active learning (machine learning) #Active perception #Human Motion and Animation #Identification (biology) #Reinforcement learning #Robot #Robot Manipulation and Learning #Robotic Locomotion and Control #Robotics #cs.RO

paper · pdf · doi:10.48550/arxiv.2605.08020

openalex publication_date 2026/05/08 · arxiv published 2026/05/08 · arxiv updated 2026/05/08 · openalex created_date 2026/05/12 · openalex updated_date 2026/07/28

Abstract

We present an active embodiment identification method for legged robots that jointly learns information-seeking behavior and explicit embodiment prediction. Using a history-augmented URMA architecture, the method infers joint-level and global embodiment parameters through interaction with the environment in simulation across different morphologies.

Citations

• Floating-Base Deep Lagrangian Networks
• LocoFormer: Generalist Locomotion via Long-context Adaptation
• Multi-Embodiment Locomotion at Scale with extreme Embodiment Randomization
• Towards Embodiment Scaling Laws in Robot Locomotion
• Gait in Eight: Efficient On-Robot Learning for Omnidirectional Quadruped Locomotion
• One Policy to Run Them All: an End-to-end Learning Approach to Multi-Embodiment Locomotion
• RL-X: A Deep Reinforcement Learning Library (not only) for RoboCup
• Human-to-Robot Imitation in the Wild
• Proximal Policy Optimization Algorithms
• Domain Randomization for Transferring Deep Neural Networks from Simulation to the Real World

Discussions

• P.S. I'm at ICRA in Vienna this week to present: Active Embodiment Identification with Reinforcement Learning for Legged Robots @ Workshop: Act to Sense to Act Better(arxiv.org/pdf/2605.08020) If you [bsky, 2 points, 0 comments]

Related

• One Policy to Run Them All: an End-to-end Learning Approach to Multi-Embodiment Locomotion
• Multi-embodiment Legged Robot Control as a Sequence Modeling Problem
• Multi-Embodiment Locomotion at Scale with extreme Embodiment Randomization
• Multi-Loco: Unifying Multi-Embodiment Legged Locomotion via Reinforcement Learning Augmented Diffusion
• Embodiment Transfer Learning for Vision-Language-Action Models