Artificial Intelligence for Regional Traffic Safety: Predicting Crash Risk and Recommending Countermeasures

Regional and remote areas carry a disproportionate share of serious road trauma, yet they are often under-served by the data-rich safety analysis methods developed for major cities. Regional roads present distinct challenges — higher speeds, longer emergency response times, mixed heavy-vehicle and local traffic, variable road conditions, and sparse or uneven crash data — that limit the direct transfer of urban safety models.

This project develops artificial intelligence and machine learning models to analyse regional traffic safety, identify high-risk locations and contributing factors, and recommend targeted, evidence-based countermeasures to reduce crash frequency and severity. Using crash records, traffic and network data, road geometry, and environmental and land-use data, the research will build predictive models of crash risk (for example, using deep learning, ensemble methods, or spatial-statistical approaches) and produce interpretable risk maps across the regional network. A key focus is not only prediction but decision support: linking identified risk factors to appropriate engineering, policy, or enforcement countermeasures, and estimating their likely safety benefit.

The outcomes aim to give road authorities such as the Queensland Department of Transport and Main Roads a practical, transferable tool for prioritising safety investment across large regional networks, contributing to the national goal of reducing road fatalities and serious injuries.