Representation learning and unsupervised skill discovery can allow robots to acquire diverse and reusable behaviors without the need for task-specific rewards. In this work, we use unsupervised reinforcement learning to learn a latent representation by maximizing the mutual information between skills and states subject to a distance constraint. Our method improves upon prior constrained skill discovery methods by replacing the latent transition maximization with a norm-matching objective. This not only results in a much a richer state space coverage compared to baseline methods, but allows the robot to learn more stable and easily controllable locomotive behaviors. We successfully deploy the learned policy on a real ANYmal quadruped robot and demonstrate that the robot can accurately reach arbitrary points of the Cartesian state space in a zero-shot manner, using only an intrinsic skill discovery and standard regularization rewards.
@article{arxiv.2410.07877,
title = {Constrained Skill Discovery: Quadruped Locomotion with Unsupervised Reinforcement Learning},
author = {Vassil Atanassov and Wanming Yu and Alexander Luis Mitchell and Mark Nicholas Finean and Ioannis Havoutis},
journal= {arXiv preprint arXiv:2410.07877},
year = {2024}
}