Hasan Al-ghanimi
Research Details
Thesis Name
Assessing the impact of AI implementation on road infrastructure success in Australia.Thesis Abstract
The continuing development and maintenance of road infrastructure plays a fundamental role in supporting Australia’s economy, public safety and community mobility (Harvey 2023). Traditional road inspection methods employed by councils and road authorities, such as LiDAR-equipped survey vehicles and manual engineering assessments, are often costly and logistically challenging, particularly in regional areas (McAnulty & Baroudi 2010). The scale and diversity of Australian road environments amplify these challenges, creating a need for more efficient and adaptable inspection methods (Mihalj et al. 2022).
Developments in Artificial Intelligence (AI) and cloud computing have initiated new opportunities for automated pavement condition inspections. Deep learning models, particularly YOLO (You Only Look Once) classification of object detection, have shown strong performance in identifying potholes, cracks and rutting using image and video data (Redmon et al. 2016)&(Fujita et al. 2023). Likewise, smartphone-based inspection methods have gained momentum because they are affordable and capable of capturing road defects data at scale (Lee, Chun & Ryu 2021). Despite these promising developments, there remains limited experiential research evaluating AI-based road inspection tools within the Australian road network particularly regarding environmental changeability, regional conditions, and alignment with national pavement classification standards.
This gap suggests a clear need for a practical and scalable AI-based solution fitted to Australia’s diverse road networks. Integrating smartphone video capture with cloud computing such as Amazon Web Services (AWS) offers significant potential for automating defect detection and reducing inspection costs across territories (Ramesh et al. 2022). However, further investigation is required to determine how precisely these systems perform compared with traditional road inspection methods, what benefits they can deliver to councils and how such platforms can contribute to long-term asset management strategies (Papke-Shields & Boyer-Wright 2017).
Therefore, this research aims to investigate the accuracy, cost efficiency and general value of an AI-based road inspection tool for Australian councils. It also seeks to explore how data generated through this system could support a future nationwide platform, enabling practical maintenance planning across regions and performing analysis of defect causes.
Research Questions:
- How accurately can AI model (AWS services and YOLO) detect and classify road defects using smartphone-capture video data?
- How does the AI-based inspection method compare with traditional road inspection practices in terms of cost and accuracy?
- What potential benefits can the proposed platform provide to councils and road authorities in terms of maintenance planning, cost reduction?
- How can the data collected from the phone be expanded into a national platform to support long-term examining of road defect causes and predictive maintenance?
