| Author: | Wu, Haitao |
| Title: | Construction quality improvement by adopting worker-robot collaboration teams and decentralized blockchains |
| Advisors: | Li, Heng (BRE) Chi, Hung-lin (BRE) |
| Degree: | Ph.D. |
| Year: | 2023 |
| Subject: | Human-robot interaction Robots, Industrial Construction industry -- Management Human engineering Hong Kong Polytechnic University -- Dissertations |
| Department: | Department of Building and Real Estate |
| Pages: | xviii, 184 pages : color illustrations |
| Language: | English |
| Abstract: | Quality is a critical metric for evaluating the value of a construction project since it directly impacts the building resilience against accidents (e.g., floods and earthquakes) and pertains to public property and lives. Unfortunately, quality failures seem to be an ever-present reality in the construction industry. Researchers have proposed various approaches for construction quality management, such as adopting quality management theories like total quality management and lean construction and implementing cutting-edge digital technologies, e.g., building information modelling, computer vision, and sensing techniques. However, the following three issues persistently impede quality performance improvement in construction practices: (1) labor-intensive construction conventions, which pose difficulties in quality control. Specifically, workers often experience fatigue due to physically demanding tasks and harsh working conditions. Fatigued workers are more prone to making mistakes, thereby degrading workmanship; (2) manual postconstruction quality inspection, which brings difficulties to effective quality control. Manual quality defect inspection (QDI) is time-consuming, subjective, and inefficient, thus reducing the reliability of quality inspection results; (3) easy-to-manipulate quality information records, which create obstacles in dispute resolution, accountability, and traceability. Quality is not determined by a single organization but by the joint work of several parties. Unfortunately, opportunistic behaviors, e.g., cutting corners and using inferior materials, are usually observed in construction collaborations, which will significantly degrade quality performance. An effective traceability system recording quality information records is required to mitigate opportunistic behaviors. The rapid development of digital technologies, especially worker-robot collaboration (WRC) and decentralized blockchains, provides creative solutions to tackle the above quality issues. WRC can integrate the robots’ advantages in strength and accuracy with human ability in intuitive decision-making and adaptability, reducing workers’ physical fatigue and minimizing quality errors. Similarly, a multi-robot system can be developed to ensure the reliability of quality inspections. Moreover, blockchain, a cryptography-based decentralized system, can meet the information management requirements for quality traceability. However, there are some gaps when utilizing these technologies. First, very few studies noticed the reliable interaction between workers and robots for safe WRC. Second, previous studies neglect the data availability and privacy in robot-based defect inspections. Third, limited attempts have been made to explore blockchain-based information management for construction process quality traceability. Finally, although blockchain seems to be a transformative tool for construction applications, we have seen very few implementations from the practices, and it is unclear related to its adoption barriers. This research aims to introduce methods to tackle these gaps and then facilitate the implementation of WRC teams and blockchains in construction quality management. Notably, this is motivated by practical industry problems rather than mere interest in new technology. The specific objectives of this research are as follows: (1) To develop a user-friendly and reliable interaction method for facilitating the transition from human-based construction to WRC; (2) To develop a multi-robot-based framework for automatic QDI; (3) To develop a blockchain-based framework for process quality traceability and accountability; and (4) To investigate barriers hindering blockchain implementation in the construction industry and identify key ones. This research first provided a comprehensive literature review on each research objective and highlighted gaps in the body knowledge. In light of these gaps, specific solutions were proposed. Specifically, this project proposed a safe and efficient method to support worker-robot interactions in WRC based on the thermal modality. An image dataset containing seven types of hand gestures was established using the thermal camera. A lightweight deep learning algorithm was developed to accurately (high accuracy) and efficiently (low latency) recognize hand gestures, even in resource-constrained mobile construction robots. Experimental results demonstrated the superiority of the proposed model compared to other lightweight algorithms and validated the feasibility of thermal image-based WRC. Subsequently, this dissertation proposed a hierarchical federated learning (FL) framework for multi-robot based QDI, allowing different construction robots to train the defect detection model collaboratively without sharing their local data. Crack detection was selected as a case study, and a lightweight segmentation algorithm was proposed to reduce communication costs. Experimental results indicated that the proposed FL method utilizes the potential of big data analysis while addressing data security and privacy concerns. After that, this thesis introduced a Hyperledger Fabric blockchain framework for extracting and recording construction process information. A consortium prototype was established using a general Blockchain as a Service (BaaS) platform. The performance was evaluated with throughput and latency metrics. Finally, this dissertation explored barriers to blockchain adoption in the construction industry, employing the technology-organization-environment (TOE) framework and identifying key obstacles through the fuzzy Decision-Making Trial and Evaluation Laboratory (DEMATEL) method. Twenty experts were invited to the survey process. Seven key barriers were identified, and corresponding policy suggestions were proposed at the government, industry, and organizational levels. This research makes contributions to the knowledge by firstly introducing a thermal image-based interaction method for safe WRC applications, exploring the potential of FL in QDI tasks, developing a blockchain framework for construction process information management, and enhancing the understanding of barriers to blockchain adoption. Moreover, the practical implications of this research include the potential to enhance quality performance by transitioning from human-based construction to WRC teams, improving the reliability of inspection results through the implementation of a robot-based QDI system, and mitigating opportunistic behaviors through blockchain-based quality traceability. |
| Rights: | All rights reserved |
| Access: | open access |
Copyright Undertaking
As a bona fide Library user, I declare that:
- I will abide by the rules and legal ordinances governing copyright regarding the use of the Database.
- 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.
- 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.
Please use this identifier to cite or link to this item:
https://theses.lib.polyu.edu.hk/handle/200/12752