Wednesday, 21 December 2022
Austroads has published practical guidance to help expand the use of machine learning (ML) and artificial intelligence (AI) in pavement asset management.
The report offers insights into how the technologies work, how to avoid common pitfalls in their development, and the types of decision making they best support.
“Machine learning and artificial intelligence are emerging technologies. They are currently predominantly used in asset management for pavement condition forecasting and optimisation of maintenance programs,” said Austroads Transport Infrastructure Program Manager, Ross Guppy. “We have looked at new and different use-cases for AI and ML, aiming to develop tools which help human experts make decisions, either by reducing the labour involved or providing additional insight to improve decisions and outcomes.”
The guidance includes a generic methodology for developing new AI/ML use-cases in pavement asset management. The application of the methodology is demonstrated in two proof-of-concept case studies, which use data from four Austroads member agencies.
The first case study sought to understand whether historical data could be used to train a ML model to reproduce expert pavement management decisions and whether the model can be applied to other road networks.
The second case study explored an extension to conventional Pavement Management System optimisation to provide insight into the network-wide implications of various multi-criteria funding allocation scenarios.
“The promising results from both case studies suggest they are feasible; however, further work is required before this technology would reach operational capability,” said Ross.
The report is accompanied by a quick reference containing a set of practical resources to help define and scope AI/ML projects. It includes advice on forming teams, recruiting specialists, managing projects, understanding key terminology and concepts, and assessing the viability and risks of specific project ideas.
“As the road sector increasingly uses data to answer enduring questions, machine learning and artificial intelligence will play an important role in future decision making. Investing in the right people, process and technologies is crucial to the success of high value ML and AI implementations.”
Download the report and quick reference: Development of Machine-Learning Decision Support Tools for Pavement Asset Management
Join us for two webinars in February 2023 with authors Tim Cross and David Rawlinson.
Success Strategies for Delivering Artificial Intelligence and Machine Learning Projects
Tuesday, 21 February 2023
This session will describe tips, tricks and success strategies to ensure AI/ML projects proceed to delivery and successful integration with existing business processes. We will work through the Quick Reference and explain how to use these tools with practical examples.
Development of Machine Learning Decision Support Tools for Pavement Asset Management
Friday, 24 February 2023
This session will describe two case studies in the use of machine learning (ML) and artificial intelligence (AI) to create decision-support tools for pavement asset management. In the first study, we explore whether ML can learn to reproduce expert treatment decisions and automatically identify candidate projects, using historic condition, treatment and inventory data. In the second, we explore extensions to conventional Pavement Management System (PMS) optimisation tools which provide insight into the network-wide implications of various multi-criteria funding allocation scenarios, and the levels of service that can potentially be realised.
No charge but registration is essential.
Can’t make the live sessions? Register and we’ll send you a link to the recording.