Skip to content

EasyPaperSniper/CrossLoco

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

9 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

CrossLoco: Human Motion Driven Control of Legged Robots via Guided Unsupervised Reinforcement Learning

CrossLoco, a guided unsupervised reinforcement learning framework that simultaneously learns robot skills and their correspondence to human motions. The key innovation is to introduce a cycle-consistency-based reward term designed to maximize the mutual information between human motions and robot states.

[paper] [website]

Annocement

Although the orginal version of CrossLoco is implemented based on LeggedGym, we now move our implementation to Orbit. Currently, this repo is still under construction but the core reward and algorithm is uploaded.

Installation

Install Orbit follow the instruction: https://isaac-orbit.github.io/orbit/source/setup/installation.html

Clone this repo under Orbit folder, then it should be fine.

Code structure

The environment and training configuration are included in ./envs

The main algorithm and runner files are locates in ./runners

About

Implementation of CrossLoco, currently lite version

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published