This repository includes the hardware and software components of "MEVION: Low-Cost Open-Source Data Collection System for Powerful and High-Speed Dual-Arm Manipulation"

Please refer to the following links for more information.

• Project Page
• arXiv
• YouTube

You can find all the hardware components in the following link: Goodle Drive

The supported environment is Ubuntu 24.04 with ROS 2 Jazzy. Install ROS 2 Jazzy Desktop before following the instructions below.

A Python virtual environment is used because Ubuntu 24.04 prevents pip from modifying the system Python environment. The --system-site-packages option is required so that the environment can also use ROS packages installed by apt and rosdep.

sudo apt update
sudo apt install python3-colcon-common-extensions python3-rosdep python3-venv

mkdir -p ~/mevion_ws/src
cd ~/mevion_ws
git clone https://github.com/haraduka/mevion.git src/mevion

source /opt/ros/jazzy/setup.bash
sudo rosdep init  # Run this only if rosdep has not been initialized before.
rosdep update
rosdep install --from-paths src --ignore-src --rosdistro jazzy -y

python3 -m venv --system-site-packages .venv
source .venv/bin/activate
python3 -m pip install --upgrade pip
python3 -m pip install -r src/mevion/requirements.txt

colcon build --symlink-install --event-handlers console_direct+
source install/setup.bash

Do not use sudo pip install or pip install --break-system-packages. Activate the ROS, Python, and workspace environments in each new terminal:

cd ~/mevion_ws
source /opt/ros/jazzy/setup.bash
source .venv/bin/activate
source install/setup.bash

# Find a description such as ATTRS{serial}=="004000325246570C20393733".
udevadm info -a -p /sys/class/net/can0 | grep serial

sudo vim /etc/udev/rules.d/99-usb-can.rules

# Add the following rule, replacing the serial number as necessary.
SUBSYSTEM=="usb", ATTRS{serial}=="004000325246570C20393733", SYMLINK+="can_usb_0"

# Unplug and reconnect the device after reloading the rules.
sudo udevadm control --reload-rules
sudo udevadm trigger

ros2 run mevion can_setup.sh

# Sense motors.
python3 -m mevion.xiaomimotor_test -i 127 -t sense

# Move motors.
python3 -m mevion.xiaomimotor_test -i 127 -t torque -v 1.0

# Change a motor ID.
python3 -m mevion.xiaomimotor_test -i 127 -t change -n 126

# Bilateral control.
python3 -m mevion.xiaomimotor_test -i 126 127 -t bilateral

python3 -m mevion.robot_utils

Make sure that the virtual environment and workspace are sourced as shown above. Start the interactive console:

ros2 run mevion ipython

Then run the following commands in IPython. mode can be single, dual, or quad. Two windows are opened: the scikit-robot viewer and MuJoCo viewer.

m = Mevion(mode="quad")
m.setup_sim()
m.robot_model.reset_manip_pose()
m.send_angle_vector(
    m.robot_model.arms.angle_vector(),
    3.0,
    interpolation="minjerk",
    comp_torque=True,
    limb="arms",
)

RViz can be launched in another sourced terminal. The robot argument accepts single_mevion, dual_mevion, or quad_mevion.

ros2 launch mevion display_launch.py robot:=quad_mevion

ros2 run mevion ipython

m = Mevion(mode="quad")
m.setup_real()
m.robot_model.reset_manip_pose()
m.send_angle_vector(
    m.robot_model.arms.angle_vector(),
    3.0,
    interpolation="minjerk",
    comp_torque=True,
    limb="arms",
)

ros2 launch mevion display_launch.py robot:=quad_mevion

• i: show axis
• w: wireframe
• o: change scale
• a: auto rotation
• s: save image
• q: quit
• f: full screen
• h: show shadow
• l: change light
• n/m: show normals

• make detailed models in solidworks and export as 3dxml
• convert 3dxml to dae using convert-urdf-mesh with --force-zero-origin option
• make simple models in solidworks and export as 3dxml
• convert 3dxml to dae using convert-urdf-mesh with --force-zero-origin option
• make simple models in solidworks and export as stl
• use converted detailed model for URDF, and use simple models for STL and DAE files
• please make mevion_dae.urdf and mevion_stl.urdf (just each file path is different)
• use mevion_dae.urdf for rviz and scikit-robot, and use mevion_stl.urdf for mujoco
• cp mevion_stl.urdf to meshes/mevion_mujoco.urdf, compile urdf into xml, change a little (stl path and geom name) and load from scene.xml

@article{kawaharazuka2026mevion,
  title={{MEVION: Low-Cost Open-Source Data Collection System for Powerful and High-Speed Dual-Arm Manipulation}},
  author={Kento Kawaharazuka and Yoshiki Obinata and Hirokazu Ishida and Jihoon Oh and Temma Suzuki and Shintaro Inoue and Keita Yoneda and Ayumu Iwata and Kei Okada},
  journal={IEEE Robotics and Automation Practice},
  year={2026},
}