EEG-based assessment in a ship's bridge simulator

• Main Author: Chia, Kimberley Ting Fang
• Other Authors: Olga Sourina
• Format: Final Year Project
• Language: English
• Published: 2016
• Subjects: DRNTU::Engineering
• Online Access: http://hdl.handle.net/10356/68322
• Institution: Nanyang Technological University

One of the most significant causes of maritime accidents is human factors. Maritime accidents are known to involve huge amount of losses and should be prevented as much as possible. Hence, the suitability of the crew is an area that will be looked into using quantitative experimental data. In this project, the study of human factors makes use Electroencephalogram (EEG). EEG can reflect the brain activity of human and it has different bands based on the frequency range, namely theta, alpha, beta and gamma. Of which, alpha band has been widely studied for its relation to relaxation and mental exertion. More specifically, the individual alpha peak frequency (iAPF) relates to cognitive abilities. The experimental procedure includes a questionnaire requiring personal information and several key factors to be studied – stress and workload levels, a series of tests for stress and workload and a simulator-based assessment. The experiments conducted aims to confirm the hypothesis of using EEG-based parameters to reflect cognitive ability. The experimental data is processed using Python and MatLab which is then analysed using Pearson’s correlation coefficient. The criterions from the questionnaire are compared against APF, individual alpha bandwidth, emotions and workload levels using correlation coefficient. Due to the limited number of subjects, there is no significant correlation obtained from the current results. However, there are several positive and negative relationships derived which confirms the hypothesis of EEG-based parameters that can reflect cognitive ability. Future work could look into the development of a software that will give conclusive results at the end of the assessment, with the help of a database of thresholds that would have been collated as part of the sample experiments. Working with a larger sample would also give better correlation between the variables and thereby, acquiring more conclusive results.