CARL: Controllable Agent with Reinforcement Learning for Quadruped Locomotion

Abstract

Motion synthesis in a dynamic environment has been a long-standing problem for character animation. Methods using motion capture data tend to scale poorly in complex environments because of their larger capturing and labeling requirement. Physics-based controllers are effective in this regard, albeit less controllable. In this paper, we present CARL, a quadruped agent that can be controlled with high-level directives and react naturally to dynamic environments. Starting with an agent that can imitate individual animation clips, we use Generative Adversarial Networks to adapt high-level controls, such as speed and heading, to action distributions that correspond to the original animations. Further fine-tuning through the deep reinforcement learning enables the agent to recover from unseen external perturbations while producing smooth transitions. It then becomes straightforward to create autonomous agents in dynamic environments by adding navigation modules over the entire process. We evaluate our approach by measuring the agent's ability to follow user control and provide a visual analysis of the generated motion to show its effectiveness.

arXiv commentProject page available at https://inventec-ai-center.github.io/projects/CARL/index.html

References in corpus (1)

• DeepMimic: Example-Guided Deep Reinforcement Learning of Physics-Based Character Skills

Cited by in corpus (11)

• AMP: Adversarial Motion Priors for Stylized Physics-Based Character Control
• ASE: Large-Scale Reusable Adversarial Skill Embeddings for Physically Simulated Characters
• ControlVAE: Model-Based Learning of Generative Controllers for Physics-Based Characters
• Learning Free Gait Transition for Quadruped Robots via Phase-Guided Controller
• GANimator: Neural Motion Synthesis from a Single Sequence
• Learning to Brachiate via Simplified Model Imitation
• Bidirectional GaitNet: A Bidirectional Prediction Model of Human Gait and Anatomical Conditions
• An Adaptable Approach to Learn Realistic Legged Locomotion without Examples
• Transition Motion Tensor: A Data-Driven Approach for Versatile and Controllable Agents in Physically Simulated Environments
• ForceGrip: Reference-Free Curriculum Learning for Realistic Grip Force Control in VR Hand Manipulation
• Deep Reinforcement Learning for High Level Character Control