Future Position Prediction for Aircraft Refuelling Port

Overview

This repository contains the code and resources for the MSc thesis “Future Position Prediction for Pressure Refuelling Port of Commercial Aircraft”. The project focuses on developing a robust framework that predicts the future position of the refuelling port on commercial aircraft, enhancing the automation of aircraft refuelling systems using deep learning techniques.

output_lkf_sa_video_lab_platform_2.mp4

Repository Structure

.
├── code/
│   ├── data_utils/
│   │   ├── data_preprocessing/
│   │   ├── dataset_analyse.ipynb
│   │   ├── dataset_preparation.ipynb
│   │   ├── synthetic_data_generation.ipynb
│   ├── framework/
│   │   ├── evaluation.py
│   │   ├── filters.py
│   │   ├── object_detection_model.py
│   │   ├── sequence_model.py
│   ├── future_position_prediction/
│   │   ├── GRU/
│   │   ├── LSTM/
│   ├── ml_studio/
│   ├── object_detection/
├── data/
│   ├── AARP/
│   ├── synthetic/
├── report/
├── requirements.txt

Summary of Main Directories

• code/

  • data_utils/: Scripts for data preprocessing, analysis, and synthetic data generation.
    • data_preprocessing/: Filtering and preprocessing scripts (e.g., EMA, Gaussian, Rolling Mean, Savitzky-Golay).
    • dataset_analyse.ipynb: Dataset analysis notebook.
    • dataset_preparation.ipynb: Dataset preparation notebook.
    • synthetic_data_generation.ipynb: Synthetic data generation notebook.
  • framework/: Core components such as evaluation metrics, filter implementations, and model definitions.
    • evaluation.py: Model performance evaluation.
    • filters.py: Data post-processing filters.
    • object_detection_model.py: Refuelling port detection logic using YOLOv10.
    • sequence_model.py: Future position prediction logic using GRU and LSTM models.
  • future_position_prediction/: Implementations for predicting future positions.
    • GRU/: GRU-based prediction models.
    • LSTM/: LSTM-based prediction models.
  • ml_studio/: Backend configurations for automating annotation processes.
  • object_detection/: YOLOv10 model configurations and training scripts.

• report/: LaTeX files for the thesis report.

Getting Started

Prerequisites

• Python 3.12.3
• PyTorch Lightning
• OpenCV
• Label Studio
• Docker
• YOLOv10

Installation

1. Clone the repository:

git clone https://github.com/AlexisBalayre/future-position-prediction-for-aircraft-refueling
cd future-position-prediction-for-aircraft-refueling

2. Create a new virtual environment using Python 3.12.3:

python3 -m venv venv

3. Activate the virtual environment:

source venv/bin/activate

4. Install the required packages:

python3 -m pip install -r requirements.txt

License

This project is licensed under the terms of the Apache V2 License. See the LICENSE file for details.

Acknowledgements

Special thanks to my supervisors, Dr. Boyu Kuang and Dr. Stuart Barnes, and sponsors Airbus, UKRI, and ATI for their support in this research through the ONEheart project.