NCTR Aircraft Classification System

Overview

The Non-Cooperative Target Recognition (NCTR) system is an AI-powered tactical display that classifies aircraft based on their flight kinematics. Using a hybrid 1D-CNN + LSTM deep learning architecture, the system analyzes trajectory data (latitude, longitude, altitude, speed, vertical rate, and heading) to differentiate between three primary classes:

• Commercial: High inertia, smooth transitions.
• Fighter: High energy, aggressive combat maneuvers.
• Drone: Low inertia, hover-capable movement patterns.

The project features a synthetic flight data generator for training and a professional Streamlit-based tactical display for real-time demonstration.

---

Project Structure

• notebooks/: Contains the model training and data generation notebook.
• web_app/: Streamlit application files (frontend and logic).
• models/: Directory for storing the trained .h5 model.
• deepkinematics/: Core logic and shared utilities.

---

Setup & Installation

1. Environment Setup

It is recommended to use a virtual environment:

python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

2. Install Dependencies

Install the required Python packages from the web_app directory:

pip install -r web_app/requirements.txt

---

Step-by-Step Execution

Phase 1: Model Generation

1. Launch Jupyter Notebook or JupyterLab.
2. Open and run all cells in notebooks/NCTR.ipynb.
3. This notebook will:
   • Generate synthetic flight trajectories for all three classes.
   • Train the hybrid CNN-LSTM model.
   • Validate performance using accuracy and confusion matrices.
   • Save the final model as deepkinematics_model.h5 in the local notebook directory.

Phase 2: Model Deployment

1. Locate the generated deepkinematics_model.h5 file.
2. Move or copy this file into the models/ directory in the root of the project.

   # On Windows (PowerShell)
   Move-Item notebooks/deepkinematics_model.h5 models/
   
   # On Linux/macOS
   mv notebooks/deepkinematics_model.h5 models/

Phase 3: Launch Tactical Display

Run the Streamlit application from the project root:

streamlit run web_app/app.py

---

Features

• Tactical 3D Plotter: Visualize aircraft trajectories in real-time.
• Intelligence Reports: Get classification results with confidence levels and probability distributions.
• Target Generation: Instantly simulate COM (Commercial), FGT (Fighter), or DRN (Drone) flight paths for analysis.
• Kinematic Telemetry: Live dashboards for Airspeed, Altitude, and Vertical Rate statistics.

Technology Stack

• Deep Learning: TensorFlow / Keras (CNN + LSTM)
• Data Science: NumPy, Pandas, Scikit-learn
• Visualization: Plotly, Matplotlib
• Web Framework: Streamlit