Fergus (Ching Yau) Mok

Research Interests

Fergus's research interests centre around the use of computer vision and data science methods in road infrastructure management. He is particularly interested in developing novel techniques to identify defects in road assets using Convolutional Neural Networks, and leveraging natural language processing techniques to extract information from road management reports. His research is focused on real-world applications, and he is collaborating with National Highways to utilize industry data in his work. 

Strategic Themes

• Entity matching in road databases: Identifying instances where a road asset is mentioned in different forms of road data 
• Information extraction and fusion: Combine different forms of data about the same asset in order to increase the utility of data available 
• Creating priority fusion list: Rank different data according to their suitability to be digitised in a road digital twin and the utility they can bring to road management 

Research Project

Enriching Digital Twins of Roads with legacy data 

Road authorities and their supply chain (road stakeholders) have historically collected and generated a large variety of data throughout the life-cycle of all existing and under construction roads. At the same time, road stakeholders are increasingly realising that there is tremendous potential value to be gained by embracing Digital Twin - based information and asset management for the entirety of a road life-cycle. Extracting usable information out of such data and fusing it in road Digital Twins is an unsolved challenge and one of the primary barriers for Digital Twins adoption in the transportation construction sector. 

Research will include: a) Devising information retrieval method(s) to extract useful information from National Highways legacy model prioritised data of A-roads; b) Applying string metrics via fuzzy matching to devise information fusion method(s) into an expressway Digital Twin; c) Combining the above methods into a data enrichment framework; d) Using additional datasets provided by the Digital Roads initiative to validate the framework via establishing the corresponding ground truth outputs and benchmarking each step’s output against its equivalent. 

This research will produce a prototype for the enrichment of expressway Digital Twins focusing on high priority information that provide the most information value to the resulting Digital Twin. This will set the foundations for encompassing existing data sources and associated processes into modern methods of working supported by Digital Twins, thus tear down one of the key adoption barriers. 

Biography

Fergus completed his Master's degree in Engineering at the University of Cambridge in 2022, with a specialization in Information and Computer Engineering, and Instrumentation and Control. His Master's project focused on utilizing lidar and RGB imagery data to automatically detect common road defects, such as potholes and cracks, using Mask RCNN. The project involved collecting data with an industry-standard lidar scanner and analysing the results generated by machine learning models that were trained on the collected dataset. 

Currently, Fergus is enrolled in a PhD program offered by the FIBE2 CDT, which is a 1+3 program consisting of one MRes year followed by three years of PhD research. He is currently in the MRes year, which includes courses that cover key research and entrepreneurial skills, in addition to technical undergraduate modules that will serve as the foundation for his research.