Controlling a mobile platform using EEG/EYE tracking and machine learning

• Main Author: Yang, Cheng
• Other Authors: Wang Lipo
• Format: Thesis-Master by Coursework
• Language: English
• Published: Nanyang Technological University 2022
• Subjects: Engineering::Electrical and electronic engineering::Control and instrumentation::Robotics
• Online Access: https://hdl.handle.net/10356/154546
• Institution: Nanyang Technological University

With the recent development of deep learning methods and equipment, the electroencephalogram (EEG) signals can be recorded and processed in various approaches. The EEG-based brain-computer interfaces (BCI) are proved to be able to handle tasks such as motor imagery, cognitive behaviors and stimulated and induced behaviors. By analyzing the voltage change of brain signals recorded by non-invasive BCI equipment and corresponding behaviors, the relationship between brain signals and behaviors can be established and further apply to practical use. In this paper, we want to use EEG to control a robotic car. The accuracy for EEG signals’ classification on task of motor imagery for 4 classes is around 0.68 to 0.75 by using different deep learning networks, therefore, eye tracking as a complementary technique is necessary to get a higher accuracy. Eye tracking is widely used in area of human vision system and psychology to analyze users’ habits. And many types of eye trackers are developed through these years which can record the position and moving speed of human pupils by looking at the stimuli. By design a scenario that each direction the eyes are looking at means one type of mission such as a robotic car turning left or right, combined with EEG signals, the accuracy to control a mobile platform will be improved. This paper mainly demonstrates an introduction to eye tracking and work of EEG classification on dataset BCI IV 2a with realization of its programs. Some exiting papers concerning combination of eye tracking and EEG are introduced as well.