TOTAL efficiency IN INTELLIGENT INFRASTRUCTURES

News

Tein Technology & Robovision : improving safety of Kennedy Tunnel with AI
02-04-2020
Alternate Text Alternate Text

In cooperation with the Flemish government, Tein Technology has launched a unique project to improve safety in the region’s tunnel network. The primary testing ground for this innovative AI approach to tunnel safety is one of the country’s most used tunnels: the Kennedy tunnel in Antwerp.

AI for a safer future on the road

The Flemish Traffic Centre, part of the Flemish Roads & Traffic Agency, called on Tein Technology to develop a system to improve tunnel safety with a ground-breaking configuration, combining a state-of-the-art camera system with additional AI analysis software.

The camera systems currently in use in the Flemish tunnel network are built with traditional technology These systems compare live imagery with predefined references of specific cases that might compromise tunnel safety. The problem is, the traditional image comparison technology cannot fully capture the continuously changing conditions inside the tunnel, setting off false alarms all too often, and missing out on less common dangerous situations.

How it works

Our goal for this project is to develop and implement an intelligent camera network, driven by AI provided by Robovision.ai. This method can accurately read even the most complex situations, drawing on millions of reference images and learning which situations may result in possible danger – and which ones most likely will not. It can effectively detect smoke, cars, trucks, pedestrians, other objects, and perfects its approach through self-learning, thus impressively enhancing response accuracy.

Testing phase complete: on to the next step

The testing phase for this technology has now been completed and the system is live on a limited number of cameras inside the Kennedy tunnel. This testing project has mapped how AI systems can be used in tunnels more efficiently. The next objective is to collect the necessary analytical results to provide a statistically relevant assessment of these systems’ scalability. In a final phase, a feasible business case will be built, taking into account all the investments – and computing power – required to roll out AI systems across the region’s entire tunnel network.


Back
Alternate Text Alternate Text