Automated Arm Project (AAP)

The Automated Arm Project (AAP) is a robotic control system built around the SO-100 robotic arm from Hugging Face's LEROBOT platform.
The 3D printable parts and hardware schematics are open-source and available on the official GitHub repository: https://github.com/huggingface/lerobot

Hardware Overview

• Robotic Arm: SO-100 by Hugging Face (3D-printed and open-hardware)
• Motors: Feetech STS3215 serial bus servos
• Controller Board: Feetech FE-URT 1
• Communication Protocol: Feetech Serial Bus Servo Protocol (half-duplex UART)

Software Overview

The project is developed in Python and controlled from the terminal.
All core scripts are located in the AAP working directory.

Current functionalities include:

• Initialization and calibration of the arm
  • Manual motor control
  • Position feedback
  • Calibration storage and reuse
• Motion recording and playback
  • Live recording of motor positions
  • CSV-based motion replay

Project Structure

• main.py : main entry point for basic motor testing and control
• main_0.py : set arm to 0 (rest) position
• main_relax.py : releases all motors (torque off)
• main_manual_control.py : manual keyboard-based motor control

utils/

• main_calibrate.py : motor manual calibration script
• port_finder.py : automatic detection of the FE-URT serial port

core/

• functions.py : utility functions for motor control, including position conversion and movement helpers

scripts/

• follow.py : movement recording and playback logic

data/

• .csv files generated by follow.py containing recorded motor positions and timestamps

config.py

• Stores motor calibration data (min, max, center) for each motor

feurt_driver.py

• Low-level communication driver for the FE-URT 1 board using the Feetech serial protocol

Roadmap

Short to Medium Term

• Improve movement precision and consistency
• Fix known bugs in main_manual_control.py
• Optimize main_relax.py (motor release routine)
• Correct CSV data parsing and replay issues in follow.py

Long Term

• Motor position curve visualization
• 3D virtual environment visualization of the robotic arm
• Integration of additional sensors:
  • Pressure sensors
  • LIDAR or ultrasonic sensors
  • Camera module

Notes

The repository is currently private but access may be granted on request.
Feedback, suggestions, and collaboration are welcome.