Harnessing rich geospatial data to transform road safety

How do we save lives on Britain’s roads when collision data is sparse and stochastic in nature?

SYSTRA’s new digital solution serves to address this issue by targeting improved multi-modal road safety across all road environments.

Our Collision Seek software, funded through a £43.5k Transport Research and Innovation Grant from the Department for Transport, provides inference to the types and severity of road collisions anywhere in Great Britain. 

This novel application uses machine learning to classify the physical environment where road collisions occur using rich geospatial data concerning features such as road geometry and markings, junction form and spacing, street furniture, pedestrian crossings, buildings and vegetation. Accordingly, highly sophisticated classifications are derived which account for nuanced changes in road character.

Aggregating collision data across similar environments

The Collision Seek tool matches locations of interest to similar physical environments nationwide where incidents have occurred. Associated STATS19 data is then aggregated across these similar environments turning sparse local records into robust, actionable risk insight.

When used across wider networks the tool can be used to evaluate the relative risk across entire road systems, identifying locations where specific users are vulnerable. This allows the end users to confidently prioritise, target and evaluate the most effective safety interventions and investment. 

Using Collision Seek to make real changes to road safety

Finding innovative solutions to projects is an important offering at SYSTRA. We want to benefit clients across a range of new projects and artificial intelligence technology forms a crucial part of the analysis. Our specialists are focused on providing the digital solutions that make transport safer, more efficient and more cost effective at every stage of development.

Road safety analysis, strategies and mitigation are undertaken at a multitude of scales. To accommodate this, we ensured that Collision Seek is adaptable for specific areas of interest, whether that be a single junction, a specific road or a whole local authority. What’s more, it also allows for mitigation that targets specific user groups or collision types. 

Developing Collision Seek to better understand road risks

Collision Seek is already being employed by our clients in delivery of an effective and pragmatic use of emerging machine learning technologies for demanded transport solutions. 

We hope it will further aid the efforts of Local Authorities, Highway Agencies and other transport bodies to save lives, reduce injuries and support decision-making for infrastructure improvements; ultimately targeting a safer and more efficient transport network. 

If you would like to have a demo or a deep dive into how it works, please reach out to Martin Campbell on [email protected]. 

You May Also LIKE