Learning a Contact-Adaptive Controller for Robust, Efficient Legged Locomotion

Abstract

We present a hierarchical framework that combines model-based control and reinforcement learning (RL) to synthesize robust controllers for a quadruped (the Unitree Laikago). The system consists of a high-level controller that learns to choose from a set of primitives in response to changes in the environment and a low-level controller that utilizes an established control method to robustly execute the primitives. Our framework learns a controller that can adapt to challenging environmental changes on the fly, including novel scenarios not seen during training. The learned controller is up to 85~percent more energy efficient and is more robust compared to baseline methods. We also deploy the controller on a physical robot without any randomization or adaptation scheme.

supplementary video: https://youtu.be/JJOmFZKpYTo

References in corpus (7)

• Learning agile and dynamic motor skills for legged robots
• Robots that can adapt like animals
• Data-Efficient Hierarchical Reinforcement Learning
• Sim-to-Real: Learning Agile Locomotion For Quadruped Robots
• GPU-Accelerated Robotic Simulation for Distributed Reinforcement Learning
• Hierarchical Reinforcement Learning for Quadruped Locomotion
• In-Hand Object Pose Tracking via Contact Feedback and GPU-Accelerated Robotic Simulation

Cited by in corpus (9)

• Learning Vision-Guided Quadrupedal Locomotion End-to-End with Cross-Modal Transformers
• Imitating and Finetuning Model Predictive Control for Robust and Symmetric Quadrupedal Locomotion
• GLiDE: Generalizable Quadrupedal Locomotion in Diverse Environments with a Centroidal Model
• Circus ANYmal: A Quadruped Learning Dexterous Manipulation with Its Limbs
• Value Iteration in Continuous Actions, States and Time
• Fast and Efficient Locomotion via Learned Gait Transitions
• Dynamics Randomization Revisited:A Case Study for Quadrupedal Locomotion
• Learning Task Space Actions for Bipedal Locomotion
• Vision-Guided Quadrupedal Locomotion in the Wild with Multi-Modal Delay Randomization