A Mobile Crowdsourcing-Based Framework for Occupant-Centric Facility Maintenance Management
Date of Award
Doctor of Philosophy
Civil and Construction Engineering
Hexu Liu, Ph.D.
Alvis Fong, Ph.D.
Yufeng Hu, Ph.D.
Osama Abudayyeh, Ph.D.
Facility maintenance management, facility management, mobile crowdsourcing, natural language processing, occupant-centric, social media
Current facility maintenance management practice relies heavily on data collection carried out by facility management professionals. At times this can result in inefficiency in facility condition data collection and decision-making for facility maintenance management. This is partially because data collection enabled by existing facility maintenance management systems lacks (1) top-down information solicitation on facility conditions, such as crowdsourcing task division, and (2) geo-referenced occupant feedback data. Mobile crowdsourcing has enormous potential to improve upon current facility maintenance management practice, especially in terms of timely data collection. In this context, this research explores the feasibility of mobile crowdsourcing for facility maintenance management and highlights the associated opportunities and challenges. A survey is conducted to understand the human, data, system, geospatial, and automation characteristics of mobile crowdsourcing for facility maintenance management on post-secondary campuses. The survey results are analyzed to reveal the challenges opportunities associated with integrating mobile crowdsourcing to achieve occupant-centric facility maintenance management, culminating in a conceptual framework for this purpose.
Furthermore, this research explores mobile social networks for facility maintenance management with a particular focus on location-specific feedback data. Specifically, a natural language processing algorithm is developed by which to automatically classify the data collected from the social network application to improve efficiency and encourage occupant participation. This algorithm can assist facility management personnel in locating and navigating failure locations and soliciting occupant feedback. Waikato Environment for Knowledge Analysis is used for training and testing machine-learning algorithms based on historical records obtained from the facility management center. The model retrieves the essential information to support facility management decision-making based on occupant input data (e.g., location and textual information). The technical feasibility of using mobile social network applications to report facility maintenance management feedback and concerns is also demonstrated using a case study.
Finally, a mobile prototype is developed to support a mobile application with location-sharing capabilities to improve upon the current system, and the process is automated using the natural language processing built in to the social media application’s chat function—customizing and training the chatbot to automatically identify the issue and speed issuing a new work order to solve the problem. In other words, this research implements natural language processing to automate responses to the end-user in human-like language in a manner that will improve communication in facility management.
This research contributes to the body of knowledge by identifying the challenges and opportunities associated with the integration of mobile crowdsourcing in facility maintenance and developing a framework for implementation. A natural language processing-based model that automatically processes the location-specific occupant feedback for facility maintenance management is developed, along with a mobile prototype that supports the creation of customizable deep neural network models to enhance communication between facility management and end-users using natural language processing. Two natural language processing techniques are integrated in one mobile application to satisfy the end-users needs.
Restricted to Campus until
Bin Alhaj, Mohamed Ahmed Madi, "A Mobile Crowdsourcing-Based Framework for Occupant-Centric Facility Maintenance Management" (2022). Dissertations. 3902.