YanjieZe/awesome-humanoid-robot-learning

A Paper List for Humanoid Robot Learning.

What it solves

This repository addresses the challenge of navigating the rapidly growing body of academic research on humanoid robot learning. It provides a curated, organized list of papers to help researchers and developers find relevant studies on how humanoid robots can learn to move, manipulate objects, and interact with their environment.

How it works

The project acts as a centralized index of academic papers, categorized by specific robotic tasks such as locomotion, manipulation, teleoperation, and state estimation. The list prioritizes research that includes real-world robot experiments and highlights entries that provide open-source code with a star (🌟) for easier implementation and verification.

Who it’s for

This collection is designed for robotics researchers, students, and engineers working in the field of humanoid robotics who want to stay updated on the latest advancements in learning-based control, whole-body motion, and simulation-to-reality transfer.

Highlights

  • Categorized by core research areas like loco-manipulation, navigation, and hardware design.
  • Prioritizes papers with validated real-world robot experiments.
  • Identifies projects with open-source code for immediate practical use.
  • Community-driven with support for pull requests to keep the list current.

Related

  • Project
    YanjieZe/GMR

    GMR is a general motion retargeting library that converts human motion data (from SMPL-X, BVH, FBX, or even monocular video) into joint commands for humanoid robots, supporting over 17 robot models and enabling real-time teleoperation and RL training data generation.

  • Project
    Roboparty/roboto_origin

    ROBOTO_ORIGIN is a fully open-source project providing the hardware designs, firmware, and software frameworks needed to build and operate a DIY humanoid robot capable of walking and running.

  • Project
    physical-superintelligence-lab/Psi0

    Ψ₀ is an open vision-language-action foundation model for humanoid robots that enables dexterous loco-manipulation by combining large-scale visual pre-training with efficient fine-tuning on robot demonstrations.

  • Project
    eProsima/Fast-DDS

    A C++ implementation of the DDS standard that provides high-performance, real-time publisher-subscriber communication for distributed systems and robotics.