Author: Fang, Xin
Title: Real-time assessment and management of mental fatigue for construction workers based on non-invasive and multimodal in-ear sensors
Advisors: Li, Heng (BRE)
Degree: Ph.D.
Year: 2025
Subject: Construction workers -- Health and hygiene
Construction workers -- China -- Hong Kong
Fatigue
Wearable technology
Hong Kong Polytechnic University -- Dissertations
Department: Department of Building and Real Estate
Pages: xxiii, 252 pages : color illustrations
Language: English
Abstract: The construction industry in Hong Kong serves as both an economic pillar and a representative high-risk industry. Construction workers, including scaffolders, equipment-operators, and welders, often have high mental workload and subsequent mental fatigue due to the harsh working environments, long working hours, and physically and mentally demanding tasks, leading to various adverse outcomes. In the short term, exposure to high mental workload and fatigue can result in decreased productivity, suboptimal work quality, and increased errors and mistakes. It also increases the risk of falls and fatalities and can impair hazard detection abilities, posing significant safety concerns. Moreover, research highlights that high mental workload and fatigue may contribute to the development of physical fatigue among construction workers. On the other hand, prolonged exposure to high mental workload and fatigue can have severe consequences on workers' health, leading to various systemic diseases such as cardiovascular disorders, stroke, depression, and other mental health illnesses. Therefore, implementing scientific measures for effective detection, management and intervention of mental fatigue among construction workers is of utmost importance for their occupational health and safety.
Given the existing deficiencies in equipment for monitoring workers' mental fatigue on construction sites, as well as the shortcomings in identifying and intervening in mental fatigue, this research aims to achieve the following objectives: (1) To develop a novel device suitable for long-term and continuous monitoring of mental fatigue among construction workers on construction sites; (2) To develop a mental fatigue classification model based on the online learning paradigm with a data filtering mechanism to enhance model learning efficiency; (3) To develop an advanced online mental fatigue classification model that considers the timeliness of physiological data flows to improve the accuracy of mental fatigue detection; (4) To develop a method for automatically generating EEG topographic maps based on the proposed multimodal in-ear device, complementing the qualitative assessment of mental fatigue; (5) To develop an ontology-based intelligent workflow for mental workload intervention, enabling the intelligent decision-making of mental fatigue management.
This research begins with a comprehensive review of previous studies on mental fatigue among construction workers, including mental fatigue detection measures, mental fatigue assessment models, and emerging technologies such as in-ear electroencephalography (EEG) devices, and summarizes the research gaps. Building upon this foundation, this research develops a novel multimodal in-ear device specifically designed for detecting mental fatigue among construction workers on construction sites. The feasibility of this device is confirmed through rigorous experimental validation. Utilizing the physiological data flows collected by this multimodal in-ear device, the fuzziness-based online sequential extreme learning machine (Fuzziness-OS-ELM) model is proposed to automatically filter the data flows, retaining samples with high data quality for improved model training and learning efficiency. Furthermore, considering the timeliness features embedded in physiological data flows, this research develops a novel regularized online sequential extreme learning machine with dynamic forgetting factor (ROSELM-DFF) model. By dynamically adjusting the weights of old and new sample data, more accurate mental fatigue detection results can be achieved. This research also introduces a novel deep learning method that enables the direct prediction of the scalp EEG signals, and generates the interpolated EEG topographic maps. Finally, this research develops an ontology-based intelligent decision-making workflow, which facilitates early warning and intervention regarding construction workers' mental fatigue.
The key findings of this research are derived from four key aspects. Firstly, the feasibility of the developed multimodal in-ear device for monitoring mental fatigue among construction workers have been demonstrated through cognitive experiments, which enables real-time and continuous acquisition and integration of electroencephalography (EEG) and electrocardiography (ECG) signals, providing a practical solution for monitoring mental fatigue in construction settings. Secondly, through multiple sets of comparative experiments, it can be concluded that samples with low fuzziness correspond to high data quality, and the proposed Fuzziness-OS-ELM model shows higher efficiency and performance in the classification of mental fatigue. In addition, this research proposes the ROSELM-DFF model, which focuses on the timeliness features inherent in physiological data flows and extracts new knowledge reflecting the mental fatigue development. Compared with batch learning-based models and other homologous models, the proposed ROSELM-DFF model offers advantages such as reduced computational complexity and exceptional mental fatigue classification accuracy, reaching an impressive 96.106%. Furthermore, this research introduces a scalp EEG prediction model and presents an interpolation method for generating precise EEG topographic maps. Through quantitative and qualitative analysis, the proposed method demonstrates its ability to accurately predict scalp EEG signals and produce highly accurate EEG topographic maps. Lastly, by enhancing the sharing and reuse of domain-specific knowledge through ontology, this research offers managers guidance and methodological framework for the effective intervention and management of workers' mental fatigue. This research has made original contributions to both theoretical and practical aspects of monitoring and intervention regarding mental fatigue among construction workers on construction sites. Theoretical contributions include innovative monitoring methods and online learning-based models that provide novel insights into the real-time assessment of mental fatigue. From the practical perspective, this research has designed and developed the specialized device and the decision-making system that enable real-time monitoring and intervention for workers' mental fatigue. These contributions have a positive impact on improving the working environment and health conditions of construction workers, as well as enhancing the safety and efficiency of the construction industry.
Rights: All rights reserved
Access: open access

Files in This Item:
File Description SizeFormat 
8030.pdfFor All Users9.03 MBAdobe PDFView/Open


Copyright Undertaking

As a bona fide Library user, I declare that:

  1. I will abide by the rules and legal ordinances governing copyright regarding the use of the Database.
  2. I will use the Database for the purpose of my research or private study only and not for circulation or further reproduction or any other purpose.
  3. I agree to indemnify and hold the University harmless from and against any loss, damage, cost, liability or expenses arising from copyright infringement or unauthorized usage.

By downloading any item(s) listed above, you acknowledge that you have read and understood the copyright undertaking as stated above, and agree to be bound by all of its terms.

Show full item record

Please use this identifier to cite or link to this item: https://theses.lib.polyu.edu.hk/handle/200/13583