Abstract:
Every year thousands of people suffer from Spinal Cord Injury (SCI) due to motor accidents, sports injuries and physical assaults. These patients experience complete or temporal changes in the function of the spinal cord caused by termination of motor neurons. Ultimately, the injury results in immobilization, loss of sensation and muscle function hence the victims have to rely on others to help them move around. This project emphasizes the need of using Brain Computer Interface (BCI) wheelchair for SCI patients to allow them to movement without relying on external aid. The idea of the proposed system includes collection of dataset, data preprocessing and classification of the Electroencephalographic (EEG) signals and then performing live control of the wheelchair. For the completion of the project data samples was collected from a 35-year-old male patient. The steps of the project include the user to think about Motor Imagery (MI) movements to collect data. EEG signals are further analyzed to find features using Fast Fourier Transform (FFT). Further these features are used as a training data to Random Forest classifier for learning and the test data will be used to measure the performance of the wheelchair. At the end, the control of the wheelchair will be performed in real time interfacing the users thought with the wheelchair. This approach is the key to reduce the probability of misclassification and improve the control accuracy of the wheelchair.
