Drone Control and System Identification using SINDy algorithm

Implements a comprehensive system for drone control, pose estimation, and system identification using the Tello drone

Project Structure

The project is organized into several main components:

1. Calibration (src/calibration/)

• camera_calibration.py: Handles camera calibration using a chessboard pattern
• Features:
  • Automatic chessboard detection
  • Intrinsic and extrinsic parameter estimation
  • Calibration data storage and loading

2. Control (src/control/)

• drone_controller.py: Main controller for the Tello drone
• Features:
  • Drone initialization and connection
  • Real-time control using keyboard inputs
  • Multi-threaded operation for control and pose estimation
  • Integration with pose estimation and camera calibration

3. Pose Estimation (src/pose_estimation/)

• pose_estimator.py: Handles ArUco marker detection and pose estimation
• Features:
  • ArUco marker detection and tracking
  • 6-DOF pose estimation
  • Real-time visualization of pose
  • Data logging for system identification

4. System Identification (src/system_identification/)

• sindy_pipeline.py: Implements Sparse Identification of Nonlinear Dynamics (SINDy)
• Features:
  • Data preprocessing and filtering
  • State and control input extraction
  • SINDy model fitting
  • Model visualization and evaluation
  • Model saving and loading

5. Utilities (src/utils/)

• keyboard_handler.py: Manages keyboard input for drone control
• Features:
  • Customizable control mapping
  • Real-time input handling
  • Takeoff and landing commands

Directory Structure

.
├── src/
│   ├── calibration/
│   ├── control/
│   ├── pose_estimation/
│   ├── system_identification/
│   └── utils/
├── sample_data/
│   ├── sample_input.csv
│   ├── sample_output.csv
│   └── sample_calibration.npz
├── calibration_data/
├── input_data/
└── output_data/

Sample Data

The repository includes sample data in the sample_data/ directory:

• sample_input.csv: Sample control input data for the drone
• sample_output.csv: Sample pose estimation data
• sample_calibration.npz: Sample camera calibration data for the Tello drone

These files can be used to test the system identification pipeline without requiring a physical drone.

Installation

1. Clone the repository:

   git clone https://github.com/Norbera0/Drone-ArUco-marker-pose-estimation-.git
   cd Drone-ArUco-marker-pose-estimation-

2. Create and activate a virtual environment:

   # On Windows
   python -m venv venv
   .\venv\Scripts\activate
   
   # On macOS/Linux
   python3 -m venv venv
   source venv/bin/activate

3. Install dependencies:

   pip install -r requirements.txt

Hardware Requirements

• Tello drone
• Computer with WiFi capability
• Chessboard pattern for calibration (9x6 grid recommended)
• ArUco markers for pose estimation

Usage

1. Camera Calibration:

   from src.calibration.camera_calibration import CameraCalibrator
   calibrator = CameraCalibrator()
   calibrator.calibrate_camera("tello")

2. Drone Control:

   from src.control.drone_controller import DroneController
   controller = DroneController()
   controller.run()

3. System Identification:

   from src.system_identification.sindy_pipeline import SINDyPipeline
   pipeline = SINDyPipeline()
   pipeline.run_pipeline("sample_data/sample_output.csv")

Control Mapping

• e: Takeoff
• q: Land
• Arrow Keys: Control yaw and altitude
• w/s: Forward/backward
• a/d: Left/right

Data Collection

The system automatically collects and stores:

• Camera calibration data
• Pose estimation data
• Control inputs
• System identification results

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

License

This project is licensed under the MIT License - see the LICENSE file for details.