Roundabout safety review using drone video analytics
This project leverages drone video analytics data collected at over 50 roundabouts in Perth to conduct comprehensive safety analyses. Building on existing footage from iMOVE project 1-028, the research will focus on analysing vehicle trajectories, speeds, and interactions with vulnerable road users.
It also aims to improve vehicle detection algorithms, implement safety surrogate measures, and develop evidence-based assessment tools for roundabout safety. By examining real-world driver behaviours and reactions to geometric design features, this work will help create more effective, proactive safety measures rather than relying solely on accident data.
Project background
While iMOVE project 1-028 (Improving roundabout modelling using drone video analytics) focused on improving roundabout modelling from a traffic engineering perspective, this project instead focuses on the safety perspective, making the best use of the already collected data.
The project addresses two main problems:
- The lack of observational data of driver behaviour
Current road safety assessment methods often rely on high-level data that summarises accident frequencies but do not provide enough detail to fully understand how unsafe driving behaviours occur. This means that treatments are often reactive – one has to wait until accidents happen before taking counter measures. To better understand the factors contributing to unsafe driving, more detailed observations are needed of how drivers react to the geometric design of roundabouts and interact with each other under varying traffic conditions.
- The absence of a roundabout safety model based on local data
Although roundabouts are designed to reduce the number of conflict points and high angle crashes compared to signalised intersections, they rely on drivers correctly yielding and negotiating the right-of-way with other drivers.
Most traffic models focus on traffic efficiency and there is no safety model that is based on WA’s local data for designers to assess safety implications. Without it, designers are forced to rely on guidelines and experience, assuming how drivers will react to certain design features.
However, current guidelines are either based on limited data from before the era of big data or derived from first principles, and they are not meant to provide off-the-shelf solutions. Therefore, it is crucial to have an evidence-based approach and develop assessment tools that are easy to use, taking into account how drivers interact under certain conditions and react to specific design features.
Project objectives
This 18-month research project aims to maximise the value of drone video data, with a renewed focus on safety analysis by reusing the same dataset.
The main objectives are to:
- Conduct vehicle trajectory and speed analysis to understand how vehicles approach and negotiate roundabouts and how they affect the opportunity to cross the roads near roundabouts for vulnerable road users (VRUs).
- Improve vehicle detection and tracking algorithms to achieve more precise measurement of vehicle dimensions and position.
- Implement safety surrogate measures that can identify risky behaviours which indicate a higher probability of incidents.
Please note …
This page will be a living record of this project. As it matures, hits milestones, etc., we’ll continue to add information, links, images, interviews and more. Watch this space!
Discover more from iMOVE Australia Cooperative Research Centre | Transport R&D
Subscribe to get the latest posts sent to your email.