haraduka

haraduka / mevius2

Public

Practical Open-Source Quadruped Robot with Sheet Metal Welding and Multimodal Perception

haraduka.github.iomevius2-hardware
47
4
100% credibility
Found Apr 01, 2026 at 47 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

Open-source hardware designs and software for MEVIUS2, a quadruped robot built with sheet metal and equipped with LiDAR, cameras, and IMU for perception and locomotion.

How It Works

1
🔍 Discover MEVIUS2

You stumble upon this exciting open-source project for building a real walking robot dog, complete with videos, designs, and research paper.

2
🛒 Gather robot parts

Download printable designs and order affordable motors, sensors, cameras, and sheet metal from the shared links to build your own quadruped.

3
🔨 Assemble the hardware

Follow simple guides to weld the lightweight metal frame, attach legs, mount sensors for seeing surroundings, and wire everything together.

4
🔌 Connect and prepare

Hook up your robot to a computer, test each motor spins correctly, and load smart walking instructions so it's ready to move.

5
🎮 Take control

Grab a game controller or keyboard to command your robot: press to sit down, stand up, or start walking around obstacles.

🚀 Robot walks!

Your custom quadruped stands tall, trots smoothly using its eyes and feelers, exploring the world just like in the demo videos.

Sign up to see the full architecture

4 more

Sign Up Free

Star Growth

See how this repo grew from 47 to 47 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is mevius2?

Mevius2 delivers a full open-source stack for building and running a quadruped robot using affordable sheet metal parts and off-the-shelf motors like RobStride03. It handles locomotion via a pre-trained reinforcement learning policy, multimodal perception with LiDAR, IMU, and camera for SLAM and elevation mapping, all in Python with ROS for real-time control over CAN buses. Developers get hardware CAD files, URDF models, MuJoCo simulation, and launch scripts to stand, walk, and navigate rough terrain out of the box.

Why is it gaining traction?

Unlike closed-source kits or simulation-only repos, Mevius2 combines practical hardware designs with deployable software, including FAST-LIO odometry and elevation mapping for real-world autonomy. The RL policy from legged_gym supports velocity commands via PS4 controller or spacenav, making it instantly testable in sim or hardware. Its sheet metal focus cuts costs while enabling custom multimodal sensing, appealing to hands-on builders.

Who should use this?

Robotics students prototyping legged locomotion, researchers tuning RL policies on physical hardware, or hardware hackers welding custom quadrupeds with LiDAR-based mapping. Ideal for those extending legged_gym to real robots or experimenting with ROS perception pipelines like FAST-LIO and elevation_mapping_cupy.

Verdict

Try it if you're into practical open-source robotics—solid docs, arXiv paper, and YouTube demos make setup straightforward despite 47 stars and 1.0% credibility score. Still early-stage with no tests, so expect tweaks for your motors or sensors.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.