SoloParkour: Constrained Reinforcement Learning for Visual Locomotion from Privileged Experience

2024/09/20 by Elliot Chane-Sane, Chane-Sane, Elliot, Joseph Amigo +7 · 7 citations
Computer Science · Neuroscience · #Computer Vision and Pattern Recognition (cs.CV) #FOS: Computer and information sciences #Machine Learning (cs.LG) #Robotics (cs.RO) #Video Surveillance and Tracking Methods #Visual Attention and Saliency Detection #Visual perception and processing mechanisms

paper · pdf · doi:10.48550/arxiv.2409.13678

openalex publication_date 2024/09/20 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

Parkour poses a significant challenge for legged robots, requiring navigation through complex environments with agility and precision based on limited sensory inputs. In this work, we introduce a novel method for training end-to-end visual policies, from depth pixels to robot control commands, to achieve agile and safe quadruped locomotion. We formulate robot parkour as a constrained reinforcement learning (RL) problem designed to maximize the emergence of agile skills within the robot's physical limits while ensuring safety. We first train a policy without vision using privileged information about the robot's surroundings. We then generate experience from this privileged policy to warm-start a sample efficient off-policy RL algorithm from depth images. This allows the robot to adapt behaviors from this privileged experience to visual locomotion while circumventing the high computational costs of RL directly from pixels. We demonstrate the effectiveness of our method on a real Solo-12 robot, showcasing its capability to perform a variety of parkour skills such as walking, climbing, leaping, and crawling.

Cited by

• Bridge the Gap: Enhancing Quadruped Locomotion with Vertical Ground Perturbations
• Autonomous Legged Mobile Manipulation for Lunar Surface Operations via Constrained Reinforcement Learning
• First Order Model-Based RL through Decoupled Backpropagation
• End-to-End Humanoid Robot Safe and Comfortable Locomotion Policy
• Constrained Reinforcement Learning for Unstable Point-Feet Bipedal Locomotion Applied to the Bolt Robot
• Quantifying and Visualizing Sim-to-Real Gaps: Physics-Guided Regularization for Reproducibility
• Bipedalism for Quadrupedal Robots: Versatile Loco-Manipulation through Risk-Adaptive Reinforcement Learning

Related

• Robot Parkour Learning
• ANYmal Parkour: Learning Agile Navigation for Quadrupedal Robots
• Parkour in the Wild: Learning a General and Extensible Agile Locomotion Policy Using Multi-expert Distillation and RL Fine-tuning
• Traversability-Aware Legged Navigation by Learning from Real-World Visual Data
• Learning to Walk in the Real World with Minimal Human Effort