Robot‑Dog STEM Kit (Open Source)

A practical, step‑by‑step quadruped robot project. It includes Fusion 360 CAD, CoppeliaSim simulation, Arduino firmware libraries, and Python tools for walking, stabilization, and tuning.

Robot-Dog

dogClip.mp4

Table of Contents

• 1. Overview
• 2. What’s in This Repo
• 3. Prerequisites
• 4. Install & Setup (Windows)
• 5. Configure COM Ports
• 6. Run Motion Control
• 7. PID Tuner UI
• 8. Setpoint Tuner UI
• 9. Simulation in CoppeliaSim
• 10. CAD: Final Design & Screenshots
• 11. Firmware Protocol (Arduino)
• 12. Calibration Data
• 13. Contributing
• 14. License
• 15. Troubleshooting

Overview

Open-source quadruped robot with:

• 3D design: Fusion 360 file 3D design/final_dog_body_design.f3d (final design)
• Simulation: CoppeliaSim/Robot_Dog.ttt
• Control: Python scripts main.py, pidTuner.py, setpointTuner.py
• Firmware libraries (Arduino/C++): libraries/cpp/…
• License: MIT (see LICENSE)

What’s in This Repo

• 3D design/ — Fusion 360 models (final: final_dog_body_design.f3d)
• CoppeliaSim/ — Scene Robot_Dog.ttt
• libraries/python/ — control modules: walk.py, pid_controller.py, serial_comm.py, etc.
• libraries/cpp/ — Arduino libraries: LegControl, MPU9250, mpu_ahrs_ekf, mpu_madgwick
• Utils/python/ — calibration, IK, stabilization, walking, Flask server examples
• Entry points: main.py (walk + stabilization), pidTuner.py (gain tuning UI), setpointTuner.py (setpoint tuning UI)
• Config: config.py (COM ports, baud, geometry, gait params)

Prerequisites

• Windows 10/11 (commands below), Python 3.12+
• Arduino IDE (install Bolder Flight Systems Eigen via Library Manager)
• CoppeliaSim (to open .ttt scene)

Install & Setup (Windows)

1. Clone or download this repo.
2. Create a virtual environment and install dependencies:

python -m venv .venv
.\.venv\Scripts\activate
pip install -r requirements.txt

3. Verify Python tools are found (optional):

python --version
pip --version

Configure COM Ports

Edit config.py and set the exact serial ports and baud:

• SERIAL_PORT_SEND — COM port that receives motor commands from Python (Arduino controlling servos)
• SERIAL_PORT_REC — COM port that streams IMU yaw/pitch/roll from Arduino
• Default baud rates: BAUD_RATE_SEND = 115200, BAUD_RATE_REC = 115200 Tip: Use Windows Device Manager → Ports (COM & LPT) to find your board’s COM numbers.

Run Motion Control

1. Connect Arduino and IMU, power servos safely.
2. Ensure your firmware implements the Firmware Protocol.
3. Start the controller:

python main.py

What it does:

• Opens two serial ports: send (motor commands) and receive (IMU)
• Starts a PID loop to compute leg deltas from IMU
• Resets leg positions, triggers a demo feature, then idles

PID Tuner UI

The UI lets you live-tune Pitch/Roll gains while the robot walks slowly. Steps:

1. Confirm COM ports in config.py.
2. Run:

python pidTuner.py

3. Use sliders/entries to adjust Kp, Ki, Kd for Pitch and Roll.

4. Observe stability and gait; lower gains if oscillations occur.

Setpoint Tuner UI

Tune desired Pitch/Roll setpoints interactively. Steps:

1. Confirm COM ports in config.py.
2. Run:

python setpointTuner.py

3. Move sliders or type exact setpoints; observe posture change.

Simulation in CoppeliaSim

1. Launch CoppeliaSim.
2. Open scene: CoppeliaSim/Robot_Dog.ttt.
3. Press Play to run the scene. Use CoppeliaSim tools to inspect kinematics. Note: Simulation is independent of Arduino; use it for planning and visualization.

CAD: Final Design & Screenshots

Find the Final Fusion 360 file at 3D design/final_dog_body_design.f3d

Top view of the robot dog body design

Side view of the robot dog body design

Firmware Protocol (Arduino)

Your firmware should:

• Stream IMU data as: <yaw,pitch,roll> at BAUD_RATE_REC (example: <0.5,-1.2,0.3>)
• Receive motor commands as:
  • Leg angles packet: <knee0,hip0,0,knee1,hip1,0,knee2,hip2,0,knee3,hip3,0> (integers)
  • Feature PWM packet (used by features.pee): <motorPWM> (e.g., <150>, and <0> to stop) Arduino Libraries:
• Install Eigen via Arduino Library Manager.
• Link (or copy) each library into Arduino’s libraries folder. Example symlink (run CMD as Administrator):

mklink /D "C:\Users\<your-user>\Documents\Arduino\libraries\LegControl" "<repo-path>\libraries\cpp\LegControl"
mklink /D "C:\Users\<your-user>\Documents\Arduino\libraries\MPU9250" "<repo-path>\libraries\cpp\MPU9250"
mklink /D "C:\Users\<your-user>\Documents\Arduino\libraries\mpu_ahrs_ekf" "<repo-path>\libraries\cpp\mpu_ahrs_ekf"
mklink /D "C:\Users\<your-user>\Documents\Arduino\libraries\mpu_madgwick" "<repo-path>\libraries\cpp\mpu_madgwick"

Upload:

• Open Arduino IDE, select your board and COM port, integrate the protocol above into your sketch, and upload.

Calibration Data

Sampling frequency: 100 Hz

• Magnetometer:

const float magn_ellipsoid_center[3] = {-16.444, 123.131, -60.0703};
const float magn_ellipsoid_transform[3][3] = {{0.887227, 0.0186386, 0.0175599}, {0.0186386, 0.878481, 0.0445495}, {0.0175599, 0.0445495, 0.978254}};

• Accelerometer & Gyroscope offsets:

#define GYRO_X_OFFSET 0.0003547
#define GYRO_Y_OFFSET 0.0002904
#define GYRO_Z_OFFSET 0.0000442
#define ACCEL_X_OFFSET -1.7587781
#define ACCEL_Y_OFFSET -8.6502850
#define ACCEL_Z_OFFSET -9.2377464

• EKF covariance (example):

const double Q[3] = {0.0000005924204051, 0.0000007749305595, 0.0000004315204792};
const double R[6] = {0.0001175380701662, 0.0001196885607162, 0.0003229845822837, 0.2635248528243216, 0.2635248528243216, 0.2635248528243216};

Reference: https://ahrs.readthedocs.io/en/latest/filters/ekf.html Note: Values above are examples. Calibrate your own hardware for best results.

Contributing

• Issues and PRs are welcome. Keep changes focused and documented.
• Include clear steps to reproduce, and tests when possible.
• Follow the project structure and existing style.

License

MIT License. See LICENSE.

Troubleshooting

• No IMU data? Confirm SERIAL_PORT_REC and run: python -m serial.tools.miniterm <COM> 115200
• No motion? Verify your firmware decodes command packets and servos are powered.
• Python can’t connect? Close Arduino Serial Monitor (only one program can open a COM port).
• Unstable gait? Lower PID gains in pidTuner.py; verify config.py geometry values.