SSVEP-based GoPiGo Navigation

This project implements a Steady-State Visual Evoked Potential (SSVEP)-based navigation system for the GoPiGo robot. The system uses EEG signal processing and machine learning techniques to classify brain signals and control the robot's movement.

Dependencies

To install the required dependencies, run the following commands:

pip install mne numpy pandas scikit-learn joblib pyriemann
pip install umap-learn matplotlib
pip install pyqt5
pip install mlxtend

How to Run

1. Install all dependencies listed above.
2. python main.pyRun the main script to start the project:

python main.py

. ├── best_model_checkpoint.h5 # Pre-trained model checkpoint ├── chrononet_live.py # Real-time ChronoNet-based EEG processing ├── ChronoNet_Model.py # ChronoNet model implementation ├── CNN_EEGNET.py # CNN-based EEGNet implementation ├── compare_and_populate.py # Script for comparing and populating results ├── comparison.py # Comparison of different models ├── config.json # Configuration file for the project ├── confusion_and_compare.py # Confusion matrix generation and model comparison ├── cortex_live.py # Real-time Cortex EEG processing ├── cortex.py # Cortex EEG processing utilities ├── CSP_SVM.py # Common Spatial Pattern (CSP) with SVM model ├── data_import_ChronoNet.py # Data import utilities for ChronoNet ├── data_import_eegnet.py # Data import utilities for EEGNet ├── data.csv # Dataset file ├── DWT_KNN.py # Discrete Wavelet Transform (DWT) with KNN model ├── DWT_SVM.py # DWT with SVM model ├── EEGNET_live.py # Real-time EEGNet-based EEG processing ├── EEGNet_Model.py # EEGNet model implementation ├── EEGNET.py # EEGNet utilities ├── Fast_Fourier.py # Fast Fourier Transform (FFT) utilities ├── FFT with PSD_SVM.py # FFT with Power Spectral Density (PSD) and SVM model ├── FFT_KNN.py # FFT with KNN model ├── FFT_new.py # Updated FFT implementation ├── FFT_SVM.py # FFT with SVM model ├── Four_class.py # Four-class classification script ├── log.txt # Log file for tracking execution ├── LSTM_Live.py # Real-time LSTM-based EEG processing ├── main.py # Main entry point for the project ├── model_svm.joblib # Pre-trained SVM model ├── oussama.h5 # Another pre-trained model checkpoint ├── predict_GRU.py # GRU-based prediction script ├── profiles.json # User profiles or configuration ├── PSD_SVM.py # PSD with SVM model └── README.md # Project documentation Key Files main.py: The main script to run the project. ChronoNet_Model.py: Contains the implementation of the ChronoNet model. EEGNet_Model.py: Contains the implementation of the EEGNet model. CSP_SVM.py: Implements CSP with SVM for EEG signal classification. FFT_SVM.py: Implements FFT with SVM for EEG signal classification. config.json: Stores configuration parameters for the project. Logs and Outputs log.txt: Contains logs generated during the execution of the project. best_model_checkpoint.h5: Stores the best model checkpoint for ChronoNet. model_svm.joblib: Stores the pre-trained SVM model. Dataset data.csv: The dataset used for training and testing the models. Authors This project was developed by Subhodeep Basu, Arko Singh, Sreerupa Roy, Bhagyashree Mane and Oussama. Contributions are welcome!