ML-based-G-Code-Attacks-Detection

This repository is associated with our manuscript (under review)

Description

This project aims to detect various types of G-Code attacks in 3D printing, such as:

• Filament Attacks: Issues like cavity creation and filament density variations.
• Thermodynamic Attacks: Manipulations involving nozzle temperature, bed temperature, and fan speed.
• Z-Profile Attacks: Alterations to the bed level of the print.

The repository includes:

• Datasets: Contains raw STL files and G-code files, datasets of benign and malicious layers and commands.
• Code: scripts for extracting features, labeling layers and commands, filtering commands, and training/testing models including LSTM, Bi-GRU, and Bi-LSTM. It also includes code for MLP and RF algorithms for command classification. Moreover, this repository includes filament attack automation scripts.

Datasets

You can access all datasets via the following link: https://drive.google.com/drive/folders/1n_4VhXoVnbwdvAuZdiioV_Eq16w2bT9O?usp=sharing

File Name Structure

-The main strcuture (Benign File): Design Name _X(dimention)_Y(dimention)_Z(height)_T(#top layers)_B(#buttom layers)_W(#walls)_L(layer thickness)_Infill Pattern_DS(infill density)_D(infill direction)_Layer Number within the file.csv

-Thermodynamic and Z_profile Attacks: Attack Class_followed by main structure.

-Cavity Attack: Attack class(filament cavity)followed by main structure Cavity_Attack_Layers (#of impacted layers)_ Lines(#of impacted lines)_length(attack magnitude)_Layer Number within the file.csv.

-Filament Speed Attack: Attack class(filament density)_followed by main structure_fd_attack_No_of_infills (#of impacted lines)_percent(Attack magnitude)_Layer Number within the file.csv.

-Filament Speed Attack: Attack class(filament state)_followed by main structure_fs_attack_No_of_infills (#of impacted lines)_Layer Number within the file.csv.

Note: Some infill patterns don't have D (infill direction).

Install the Environment

# Using Pyothn:
python -m venv myenv
source myenv/bin/activate

# Using Miniconda
1. Download the Miniconda installer
wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh

2.Make the installer script executable
chmod +x Miniconda3-latest-Linux-x86_64.sh

3.Run the installer script
./Miniconda3-latest-Linux-x86_64.sh

4. Activate the Miniconda environment
source ~/.bashrc

5.Verify the installation
conda --version

6.Create and activate a new Conda environment
conda create --name myenv python=3.9.18
conda activate myenv

# Install TensorFlow and other necessary packages
pip install tensorflow[cuda]
conda install scikit-learn keras pandas

#Checking GPU Availability (Optional)
# Verify that TensorFlow can access the GPU
python3 -c "import tensorflow as tf; print(tf.config.list_physical_devices('GPU'))"

#Note:  Ensure that both TensorFlow and CUDA versions are compatible.