简体中文 | English
Chinese and English learning materials for a future Foundation of Embodied AI course, based on our project experience and generated with the assistance of ChatGPT.
How to read: Follow tools → calibration/deep learning → physics/perception → reinforcement learning → demonstrations/generative policies → systems integration. Git and ROS 2 are optional tool references. See the course map · PDF.
17 topics, 34 Markdown files and 34 PDFs. Use the links below to read or download. All PDFs use a shared academic layout; the first chapter starts on page 2. Includes VLA imitation learning, FQI/SAC, and Dream-1 zero-shot embodiment transfer.
| No. | Topic | Chinese Markdown | English Markdown | Chinese PDF | English PDF |
|---|---|---|---|---|---|
| 1 | Git Basics | MD | MD | ||
| 2 | ROS 2 Basics | MD | MD | ||
| 3 | Camera Calibration | MD | MD | ||
| 4 | Deep Learning Basics | MD | MD | ||
| 5 | Vision Foundation Models | MD | MD | ||
| 6 | Robot Dynamics | MD | MD | ||
| 7 | Isaac / MuJoCo Simulation | MD | MD | ||
| 8 | Motion Control RL | MD | MD | ||
| 9 | PPO / Actor–Critic | MD | MD | ||
| 10 | Q-learning / SAC | MD | MD | ||
| 11 | Imitation Learning | MD | MD | ||
| 12 | Flow Matching | MD | MD | ||
| 13 | GR00T & SONIC | MD | MD | ||
| 14 | World Action Models | MD | MD | ||
| 15 | Efficient Training & Inference | MD | MD | ||
| 16 | Post-training | MD | MD | ||
| 17 | Robot Agents | MD | MD |
Everyone is welcome to download, modify and submit pull requests under the applicable license. Send feedback or requests for separate commercial permission to zengx3@sustech.edu.cn.
New original contributions: From 2026-10-22, original additions first published after the license boundary commit use CC BY-NC-SA 4.0, where the licensors hold the required rights:
- Attribution: Credit the creators, link to the source and license, and identify changes. Do not imply endorsement.
- NonCommercial: Commercial use is not authorized by this license; request separate permission from the relevant rights holders.
- ShareAlike: When sharing adaptations of these contributions, use CC BY-NC-SA 4.0 or a license officially recognized as compatible.