Click-and-Traverse
- AI model
- Navigation and mapping
A reinforcement learning framework for teaching humanoid robots to walk through cluttered indoor scenes without colliding with obstacles, built on MuJoCo simulation with Unitree G1 assets.
Made by Galbot.
Source on GitHub: github.com/GalaxyGeneralRobotics/Click-and-Traverse (opens in a new tab). Facts read at commit 866ba39 (opens in a new tab).
Licenses
- Code
- Apache-2.0 (opens in a new tab)
Read by Robotheca from LICENSE (opens in a new tab) (Apache-2.0) at commit 866ba39 (opens in a new tab), on 16 Oct 2026.
- Weights
- Apache-2.0 (opens in a new tab)
Read on huggingface.co/Axian12138/Click-and-Traverse (opens in a new tab).
- Commercial use
- Yes
What it does
- Encodes spatial relations between the humanoid body and nearby obstacles as a potential-field representation called HumanoidPF.
- Generates hybrid obstacle scenes, combining realistic indoor crops with procedurally synthesized typical and random obstacles.
- Trains specialist policies on specific scenes with reinforcement learning, then distills them into a generalist policy via DAgger.
- Provides a generalist v1 checkpoint for community testing, with ONNX export and batch success-rate evaluation tools.
- Includes real-world deployment code in a separate SLAM-based deployment directory.
Written by Robotheca from the project's README (opens in a new tab) at commit 866ba39.
Getting started
- Difficulty
- Advanced
- How to start
- The project's website (opens in a new tab)
Maintenance
- Status
- Maintained: a commit in the last 12 months
- Newest commit
- 7 Jul 2026 on main
- Latest release
- None published on GitHub
Checked on GitHub on 16 Oct 2026.
Links
Robots that can run it
No robot on Robotheca is linked to it yet. Links come from a robot's robot.yaml, from its dependency files read as text, or from Robotheca's editors.
Robotheca lists AI models as a reference: nothing is hosted or run here, and every fact links to where it was read.