A European national operator managed 3,000 miles of motorway in on the Europe’s largest economies.

This motorway network caries 30% of the nation’s traffic, is monitored by 7 regional operating centers, 30,000 edge devices and 1,000 patrol vehicles. These represent an annual investment of > $300M / year in monitoring and management of its network. The operator suffered many incidents connected to hazards and stopped vehicles and issued a challenge to use technology to improve detection and response to road hazards.

The operator tried many types of hazard perception technologies, but found that the complexity, size, and ITS coverage variety on its network meant that a single type of hazard perception was not sufficient. Despite significant investment in ITS, the operator found that it was receiving thousands of alerts / day, the majority of which were either not true or not actionable.

"Wow, this is impressive; much more than what I expected. It enables efficiency as we can detect more, and more accurately.”
– Director of Operations, European network operator

Within 8 weeks of being deployed, Lanternn’s AI was able to significantly improve both detection quality and detection time of incidents. This is especially potent considering the huge investment the operator makes every year into detection and response, showcasing an ability to significantly improve on existing operations using Valerann’s AI, or alternatively enjoy significant savings while maintaining similar levels of service. The system is now continuing into operational use on some the Nation’s most complex infrastructure.

The operator together with Valerann deployed Lanternn by Valerann™ on a 600-mile section of its network in the center of the country.

The system monitored 5 interconnecting motorways, as well as a ring road around the nation’s 2nd largest city. The system integrated with 120 CCTV cameras, over 300 loops, Waze, HERE, INRIX, TomTom and Google maps, as well as weather data. Valerann committed to a tight timeline and delivered this complex system with all the integrations within

‍
8 weeks. The system used machine vision, data fusion and AI to sift through over 750,000 data points a day and surface only the 600 most relevant events to the control center. The alerts were coupled with contextual data and historical video footage for rapid validation. The system was compared to other incident detection capabilities, to estimate both accuracy and ability to detect novel incidents.