AI/ML to detect road safety issues in real-time

Transoft Solutions technology to form backbone to Ontario crash risk prediction system

Transoft Solutions is working in partnership with the Autonomous Vehicle Innovation Network (AVIN) in Ontario, Canada to develop an AI-based crash risk prediction system, which includes near-miss notification alerts to improve traffic safety across the regional municipality of Durham.

The underlying technology uses artificial intelligence and deep learning to automate video analysis of traffic flow.

Based in Waterloo, Ontario, Transoft Solutions, a developer of transportation engineering software and road safety analytic technology, has matched the $876,000 in support from Ontario’s AVIN R&D Partnership Fund with $1,798,350 in industry funding for a $2.67M project designed to tackle collisions and everyday road safety issues in a proactive way.

Annually, tens of millions of people around the world are injured in traffic collisions. Cities have historically relied on collision data—collected after vehicle impacts have occurred—to determine the most dangerous roads and decide which interventions might reduce injuries and fatalities.

However, that data, compiled over a period of years, rarely details all the factors contributing to collisions. And while many cities have recently installed cameras on streets and at intersections, manually checking countless hours of video footage is a daunting task.

The AVIN crash risk prediction project will extend the core capabilities of Transoft’s automated road safety solutions to allow collision courses between road users to be predicted in real-time so that potential crashes can be detected in advance and avoided. And to further enhance the prediction accuracy and computation efficiency, AI will be moved from the cloud to the edge to accelerate advanced collision warning time for users.

Additionally, by monitoring and understanding the patterns of these near-misses, a system for forecasting long-term safety problems will also be included. This data informs planners when allocating road safety budgets and solutions.

Transoft’s automated road safety solutions use artificial intelligence and deep learning technology to automate video analysis of traffic flow. The technology detects crashes, near-misses, and other potential safety concerns, such as excessive speeding. This empowers transportation engineers and planners to better understand traffic patterns and road user behaviour to make informed decisions that improve safety before a crisis occurs.

“Traffic safety technologies are going through an exciting transformation today. We at Transoft Solutions are proud to receive the support from the Ontario government to expedite the commercialisation of our innovative technologies that reduce road fatalities. This project will further enhance our leadership position in the industry while making Canadian roads safer,” said Charles Chung, VP Transportation Safety at Transoft Solutions.

“The Regional Municipality of Durham has partnered with Transoft Solutions to help look for new ways to improve safety at intersections. Durham Vision Zero has identified that a majority of collisions occur at intersections and in order to reach our goal, new innovative and effective methods are necessary.  Real-time prediction and crash forecasting is an example of this, which diagnoses safety issues prior to collisions occurring. The Region of Durham is thrilled to be part of this progressive team,” said Amanda Spencer, Project Manager Road Safety Group at the Regional Municipality of Durham.

The post AI/ML to detect road safety issues in real-time appeared first on AEC Magazine.