Crowd-sourced data use in generating road safety insights

The Using crowdsourced data to improve road management project, with iMOVE partners Department of Transport and Main Roads (Queensland), Queensland University of Technology, and CARRS-Q has been completed. Its final report is available for download below.

The use of crowd-sourcing in transportation systems, particularly in road management and safety applications, is a growing area of interest. Traditional methods of collecting traffic data and road safety information, such as manual surveys, camera-based systems, and fixed sensors, often suffer from high costs, limited coverage, and slower data refresh rates.

In contrast, crowdsourcing provides a more scalable and near real-time approach to gathering such data by leveraging the power of widespread participation through mobile devices, connected vehicles, and advanced sensor technologies, using existing Advance Driver Assistance Systems (ADAS).

This project focused on utilising crowd-sourced data for generating road safety insights, aimed to understand the capabilities of such data by developing two prototype applications for road infrastructure managers.

The first application developed an automated process to produce MAP Extended Messages (MAPEM) in an intersection. MAPEM is a Cooperative Intelligent Transport System (C-ITS) message with detailed road topology information used by the Road and Lane Topology (RLT) service.

The second application endeavoured to produce as many parameters as possible from the International Road Assessment Program (iRAP). During the project delivery, Multi-Modal Large Language Models (M-LLM) became available and therefore, the project team also demonstrated some iRAP attributes can be extracted using latest M-LLMs simply by using images of the road network. The project also assessed the crowd-sourced data quality

Expected project impacts

Crowd-sourced data, especially from Advanced Driver Assistance Systems (ADAS), offers new approaches to generate road safety insights. As Multimodal Large Language Models (M-LLMs) emerge, the potential applications for infrastructure management and safety analysis will only continue to grow.

Professor Sebastien Glaser

This project demonstrates the abundance of valuable information now accessible through crowd-sourced data. However, effectively utilising this data will still necessitate validation by skilled professionals.

Amit Trivedi, Director, Safety Technology Products, Department of Transport and Main Roads

Download the final report

Download your copy of the final report, Using crowdsourced data to improve road management, by clicking the button below.