Super Brain

Super Brain is a Brain-Controlled Wheelchair (BCW) prototype that helps people with severe physical disabilities move independently using their brain signals.

Instead of using a joystick or physical controls, the system reads EEG (electroencephalography) signals, processes them, and maps them to movement commands such as forward, backward, and stop.

1. Project Overview

The main idea of this project is:

• Capture raw EEG signals from a brain-computer interface (BCI) headset.
• Clean and process these signals so that they are usable for a machine learning model.
• Train a classifier that can recognize different mental states or intentions.
• Convert the model’s predictions into wheelchair navigation commands.

This repository currently focuses mainly on the data and analysis side of the system: preprocessed EEG datasets and helper scripts for exploring and comparing them.

2. Key Features

• Brain-Computer Interface (BCI)
  Uses EEG signals as input instead of physical buttons or joystick.

• Movement Classification
  Supports different navigation intents such as:

  • Move forward
  • Move backward
  • Stop / no movement

• Data-Driven Approach
  EEG samples are stored in CSV files, labeled by the type of movement, and used to train and evaluate models.

• Modular Design
  Data, scripts, and project files are separated so it is easy to extend the system with new commands, models, or sensors.

• Assistive Technology Focus
  Designed with accessibility in mind to improve quality of life for users with limited motor control.

3. Repository Structure

At a high level, the repository contains:

• Project/
  Contains the main project code and/or notebooks used for:

  • Loading EEG datasets
  • Cleaning and preprocessing data
  • Training and evaluating machine learning models
  • Experimenting with different approaches for classification

• Final.csv
  A consolidated dataset of EEG features and labels, typically used for final training or evaluation.

• Final_forward.csv
  EEG samples corresponding to forward movement intent.

• Final_backward.csv
  EEG samples corresponding to backward movement intent.

• stop.csv
  EEG samples representing stop / neutral / no movement intent.

• datadiff.py
  A Python utility script for working with the data. Typical operations you would expect from this script include:

  • Comparing different CSV files (e.g., forward vs. backward vs. stop)
  • Computing basic statistics or differences between datasets
  • Helping in feature exploration / sanity checks

• .codebuddy/ and .vscode/
  Editor and tooling configuration for a smoother development experience.

• .gitignore.txt
  Lists files and folders that should be ignored by Git (temporary files, environment files, etc.).

4. How the System Works (Conceptual Flow)

1. EEG Signal Acquisition

   • EEG headset captures brain signals while the user performs certain mental tasks (e.g., thinking “move forward”).
   • Signals are recorded and saved into CSV files with labels such as forward, backward, or stop.

2. Preprocessing & Feature Extraction

   • Noise is reduced (for example, via filtering and artifact removal).
   • Signals may be segmented into time windows.
   • Features (like power in certain frequency bands, statistical features, etc.) are extracted and written into CSV files:
     • Final_forward.csv
     • Final_backward.csv
     • stop.csv
   • These may later be merged into Final.csv.

3. Model Training

   • The combined dataset (Final.csv) is loaded into a machine learning pipeline.
   • Data is split into training and testing sets.
   • A classifier (for example: SVM, Random Forest, or Neural Network) is trained to distinguish between different commands based on EEG features.

4. Prediction & Command Generation

   • In a real-time system, incoming EEG data would be preprocessed and passed to the trained model.
   • The model outputs a predicted class: forward, backward, or stop.
   • This prediction is then translated into a wheelchair command (e.g., send a signal to the wheelchair controller).

5. Safety Layer (Conceptual)

   • Safety checks (e.g., confirmation, time thresholds, or additional sensors) can be added so that the wheelchair only moves when the prediction is confident and safe.

5. Getting Started

5.1. Prerequisites

You will need:

• Python 3.x
• Common data/ML libraries such as:
  • pandas
  • numpy
  • scikit-learn
  • matplotlib (or any plotting library)

Install them using:

pip install pandas numpy scikit-learn matplotlib