The Mini Arm is a miniature version of the Desktop-Arm project, a portable 6DOF 3D-printed open-source robot arm. Plug in a USB cable and immediately enjoy these features:
- Runs on Raspberry Pico 2 microcontroller using CircuitPython
- Internal IK solver to handle joint state calculations on the microcontroller!
- 95% 3D printable components (aside from nuts & bolts)
- Total parts cost <= $100
- Weighs less than 1lb (~0.3kg)
These install instructions were tailored using a Windows 10 OS. There are also setup instructions for Ubuntu and the Raspberry Pi too.
- Repository Structure
- Installation
- Demos
- Tests
- Multimodal agentic AI Integration
- AJ GUI
- ROS2 Robot Visualizer
- Acknowledgements
This repository contains everything needed to build, program, and control the Mini-Arm robot:
Mini-Arm/
├── examples/ # Usage examples and demos
│ ├── 01_basic_control/ # Simple control scripts
│ ├── 02_trajectory/ # Trajectory execution
│ ├── 02_xbox_teleop/ # Xbox controller teleop
│ └── 03_analysis/ # Motion analysis tools
├── tests/ # Test scripts organized by category
│ ├── hardware/ # Serial & servo tests
│ ├── kinematics/ # IK solver tests
│ ├── networking/ # Network communication tests
│ └── visualization/ # 3D visualization tests
├── pico/ # CircuitPython firmware for Pico
├── miniarm_ros/ # ROS2 packages for visualization
├── assets/ # Images, models, documentation
├── mini_arm.py # Python client library
└── requirements.txt # Python dependencies
Key files:
mini_arm.py- MiniArmClient class for serial communicationpico/code.py- Main firmware running on Raspberry Pi Picorequirements.txt- Python package dependencies
-
Choose Pico firmware
We have provided CircuitPython firmware for both the Pico and Pico2! Choose the right uf2 file for your Pico version, or download the CircuitPython latest firmware for the Pico or Pico 2.
-
Move/drag the uf2 file to the Pico drive.
- When you plug in the Pico for the first time, you should see a new drive appear on your computer with the Poci in bootloader mode. You can always access this by pressing the BOOTSEL button while plugging in the Pico.
- Move/drag the uf2 file to the Pico drive. The Pico will automatically reboot and the LED will turn green to indicate successful flashing.
- When you plug in the Pico for the first time, you should see a new drive appear on your computer with the Poci in bootloader mode. You can always access this by pressing the BOOTSEL button while plugging in the Pico.
-
Replace the contents of the
code.pyandlibfolder to the Pico drive.- The
code.pyfile contains the main code to run the Mini Arm. Open this file to enable/disable certain settings.
- The
-
Building the arm
- The full list of parts needed to assemble the arm can be found in the B.O.M.
- Files for 3D printing can be found on HowToMechatronics with assembly instructions
-
Electrical wiring
The mini_arm.py module provides a simple CLI:
# Get list of available commands
python mini_arm.py --port COM3
# Send a specific command
python mini_arm.py --port COM3 --command "get_pose"
# Interactive mode
python mini_arm.py --port COM3 --interactiveUse the MiniArmClient class in your own scripts:
from mini_arm import MiniArmClient
# Connect to Mini-Arm
client = MiniArmClient(port='COM3', baudrate=9600, verbose=True)
# Send commands
client.send('help') # List available commands
client.send('home') # Move to home position
client.send('get_pose') # Get current position
client.send('set_pose:[0.135,0.0,0.22]') # Move to positionThe Mini-Arm supports the following commands:
================================= List of commands =============================================
movemotor | MOTOR VALUE | // Moves motor A to absolute position B (deg)
movemotors | VALUES | // Moves motors absolute position B (deg) assimung VALUES is a list
info | | // Prints info about robot system (motors, grippers, and sensors)
set_gripper | VALUE, STRING | // Gripper command to set the state (open/close) or position
set_pose | VALUES | // Updates the end effector pose to an absolute cartesian coordinate pose. Pass a list of values (ex: [X,Y,Z] or [X,Y,Z,R,P,Y])
get_pose | | // Returns the current position and orientation of the robot end effector
set_joints | (DISABLED) | // Updates the robot joint state. Pass a list of values (ex: [0,0,0,0,0,0])
get_joints | | // Returns the current robot joint state
set_led | VALUES | // Set the RGB LED to a specific color using 0-255 values in a 3-element list
set_delta | VALUES | // Updates the end effector pose with a delta movement relative to the robot's current pose. Pass a list of values (ex: [X,Y,Z] or [X,Y,Z,R,P,Y])
posture | (DISABLED) | // Updates the end effector pose with a Cartesian displacement relative to the robot's origin. Pass a list of values (ex: [X,Y,Z] or [X,Y,Z,R,P,Y])
controller | (DISABLED) | // Controller-specific message as a long ascii string with buttons and joystick data that gets converted into a delta position. Check 'xbox_utils' for message type details
help | | // Display available commands
play_music | STRING | // Play a music file
debug | STRING | // Pass 'on' or 'off' to enable or disable the verbose output
set_rate | VALUE | // Update the main loop rate (Hz)
trajectory | STRING [REPEAT] | // Perform a specified trajectory (e.g. 'circle') with optional repeat argument (true/false)
stop | | // Stop ongoing traetories
fsr | | // Read the sensor values from the FSRs
home | | // Set the robot to its home position
test | | // Test command to verify output from device
================================================================================================
See the examples/ directory for detailed usage examples:
basic_demo.py- Connect and send simple commandsposition_control.py- Move end effector to XYZ coordinatesgripper_control.py- Open/close gripper
circle_trajectory.py- Execute circular trajectories in 3D space
Control the arm with an Xbox controller in real-time:
cd examples/02_xbox_teleop
python xbox-client.py --port COM3Features:
- Real-time end effector control via joysticks
- Gripper control with triggers
- LED status indicators (green = connected)
trajectory_command.py- Generate and execute trajectoriestrajectory_comparison.py- Compare commanded vs actual motionconvert_c3d_to_csv.py- Process motion capture datacompute_alignment_transform.py- Align coordinate frames
Test scripts are organized by functionality in the tests/ directory:
- hardware/ - Serial communication and servo control tests
- kinematics/ - Inverse kinematics solver validation
- networking/ - TCP/IP and camera streaming tests
- visualization/ - 3D rendering and simulation tests
See tests/README.md for details on running tests.
TBD
Robot control cana also be done through the Unity version of the "AJ" GUI v1 (Link coming soon), or an SSH terminal to the robot. FOR TCP/IP, quickest way using a python terminal:
The Mini Arm can be visualized in RViz2 using the included ROS2 packages. Refer to the build instructions for more information.
Feel free to reach out to me in case of any issues.
If you find this repo useful in any way please do star ⭐️ it so that others can reap it's benefits as well!
This project is inspired from the work done by:
Copyright 2022-2023 Jonathan Shulgach
This Source Code Form is subject to the terms of the Mozilla Public License, v. 2.0. If a copy of the MPL was not distributed with this file, you can obtain one at https://mozilla.org/MPL/2.0/.