aiqbench analysis
This research addresses the critical sim-to-real gap in legged robotics by leveraging Nvidia's Isaac Sim and Lab to train reinforcement learning policies that transfer directly to physical hardware without fine-tuning. Validated on a Unitree Go1 quadruped, the framework demonstrates superior disturbance recovery compared to standard integrated controllers while maintaining high-speed locomotion capabilities. The work highlights how advanced simulation environments are accelerating practical deployments of complex whole-body control systems in real-world scenarios.
Isaac Sim-to-Real: Reinforcement Learning based Locomotion for Quadrupeds
arXiv:2607.18135v1 Announce Type: cross Abstract: Learning-based approaches to locomotion have risen in popularity in recent years, showing the capability for complex legged locomotion and whole-body control. Reinforcement learning (RL), the primary learning-based approach for locomotion, often utilizes a high-performance simulation tool, providing a controlled and efficient training and development environment. However, policies that perform well in simulation frequently encounter unexpected challenges when deployed on a physical system, known as the sim-to-real gap. This work presents a robust RL locomotion framework capable of whole-body control. The proposed RL framework utilizes Nvidia's new set of simulation tools, Isaac Sim, and its companion RL framework, Isaac Lab, for training, achieving a zero-shot sim-to-real policy. The performance of our policy is validated on physical hardware using the Unitree Go1, with experimental results showing similar velocity tracking performance to the quadruped's integrated controller, with a greater ability to recover from large disturbances, and achieve linear velocities of 2.0 m/s and angular velocities of 1.8 rad/s.
Abstract:Learning-based approaches to locomotion have risen in popularity in recent years, showing the capability for complex legged locomotion and whole-body control. Reinforcement learning (RL), the primary learning-based approach for locomotion, often utilizes a high-performance simulation tool, providing a controlled and efficient training and development environment. However, policies that perform well in simulation frequently encounter unexpected challenges when deployed on a physical system, known as the sim-to-real gap. This work presents a robust RL locomotion framework capable of whole-body control. The proposed RL framework utilizes Nvidia's new set of simulation tools, Isaac Sim, and its companion RL framework, Isaac Lab, for training, achieving a zero-shot sim-to-real policy. The performance of our policy is validated on physical hardware using the Unitree Go1, with experimental results showing similar velocity tracking performance to the quadruped's integrated controller, with a greater ability to recover from large disturbances, and achieve linear velocities of 2.0 m/s and angular velocities of 1.8 rad/s.
6 pages, 5 figures. Accepted manuscript. Published in the 2025 IEEE 21st International Conference on Automation Science and Engineering (CASE), pp. 2194-2199
Robotics (cs.RO)
arXiv:2607.18135 [cs.RO]
(or arXiv:2607.18135v1 [cs.RO] for this version)
https://doi.org/10.48550/arXiv.2607.18135 arXiv-issued DOI via DataCite (pending registration)
2025 IEEE 21st International Conference on Automation Science and Engineering (CASE), pp. 2194-2199 (2025)
https://doi.org/10.1109/CASE58245.2025.11163761 DOI(s) linking to related resources
Full text under the publisher's open licence · 214 words
Related coverage
The State of Simulation for Physical AI: An Overview
Hugging Face · Jul 21, 2026
Scale Robot Reinforcement Learning with NVIDIA Isaac Lab on Amazon SageMaker AI
AWS · Jun 9, 2026
Energy-Efficient Gait Adaptation via Hierarchical Reinforcement Learning for Quadrupedal Locomotion Across Diverse Terrains
arXiv · Oct 8, 2026
Neuromorphic Reinforcement Learning for Quadruped Locomotion Control on Uneven Terrain
arXiv · Jun 10, 2026