Mental stress detection for vessel traffic service operators based on EEG data

• Main Author: Wen, Justin JiXi
• Other Authors: Chen Chun-Hsien
• Format: Final Year Project
• Language: English
• Published: Nanyang Technological University 2023
• Subjects: Engineering::Mechanical engineering
• Online Access: https://hdl.handle.net/10356/168080
• Institution: Nanyang Technological University

Vessel Traffic Service Operators (VTSOs) are responsible for ensuring the safe and efficient operation of waterways. They use a Vessel Traffic Management System (VTMS) to provide real-time information and ensure the smooth flow of vessel traffic. However, the demanding nature of their work can lead to stress, which can impact their performance and lead to physical and mental health problems. To alleviate stress on VTSOs and reduce the risk of maritime accidents, investment in stress management technologies is important. In this study, the proposed solution is a novel machine learning model called 3D Mixture of Experts Convolutional Neural Network (3DMoEConvNet) to predict stress level for VTSOs based on their EEG signals. The 3DMoEConvNet model combines the strengths of 3D Convolutional Neural Network (3D CNN) and Mixture of Experts (MoE) architecture to effectively capture both the spatial and temporal features of EEG data, while also addressing the issue of individual differences in EEG data. The 3DMoEConvNet achieved accuracies of 99.77%, 99.84% and 99.83% for 2-Class, 3-Class and 4-Class predictions respectively. The proposed model provides a basis for the advancement of EEG-based stress detection systems.