Deep Learning for Brain Computer Instrumentation

This study presents a novel application of a 1-Dimensional Convolutional Neural Network (1D-CNN) model in converting electroencephalography (EEG) signals into drone control commands. The model is designed to recognize and classify six distinct movement commands - backward, forward, left, right, land, and takeoff - by identifying specific brain wave patterns. The EEG dataset, comprising 2,498,750 records, was collected using a 16-channel EEG headset with a sampling rate of 125 Hz. Following data preprocessing and feature extraction, the proposed 1D-CNN model was trained and achieved an accuracy of 99.27% in categorizing brain wave patterns based on commands. This research contributes to the growing field of brain-computer interfaces (BCIs) and offers promising prospects for the development of EEG-controlled drones.