Inside the A75 One Highway. Constantly Changing Conditions.
Nearly two decades of intelligent traffic monitoring on one of France’s most demanding highways
A road does not behave the same way all day, let alone all year. Light changes, traffic builds and thins, rain arrives, visibility drops, and incidents have an inconvenient habit of happening at exactly the wrong moment.
That becomes particularly noticeable as autumn moves into winter across the Northern Hemisphere. With the Earth tilted away from the Sun, daylight hours shorten and the Sun sits lower in the sky. For drivers, that means more journeys in darkness and periods of low, awkward light around dawn and dusk. For the people responsible for keeping roads moving, it means the scene they are monitoring is constantly changing.
The A75 in France is a useful example of what that looks like in practice. France’s north-south highway passes through a varied landscape and includes five tunnels, as well as sections where long sight lines, bends and bidirectional traffic create quite different monitoring challenges. Its intelligent transport system has been developing since 2007, with Teledyne FLIR's technology used across tunnels and critical highway sections and two control centers supporting incident detection and supervision.
A Road With More Than One Personality
A tunnel is a controlled environment in some respects, but not an easy one. Vehicles move rapidly between daylight and artificial light, traffic can travel in different directions, and an incident in an enclosed space can become serious quickly. An open section of highway presents another set of problems: distance, weather, changing light and a much larger field of view.
The A75 system therefore does not rely on one approach everywhere. The project includes TrafiBot cameras, ITS Dual systems and Cascade software, with different configurations used across tunnels and highway sections. One section, at Virage de Coudes, uses 36 TrafiBot Dual AI cameras for long-distance detection and bidirectional traffic monitoring.
For the operator sitting in a control room, however, the technical detail ultimately has to resolve into something much simpler: what is happening on the road, and does somebody need to do something about it?
Automatic incident detection helps answer the first part. Video analytics can identify situations such as stopped vehicles or unexpected objects and bring them to an operator’s attention. The operator can then look at the live images, understand the context and decide what response is appropriate.
It is a good example of where artificial intelligence is most useful when it is not trying to replace the person in the control room. Its job is to keep looking.
When the View Changes
The seasonal theme matters here because road monitoring is fundamentally a visibility problem.
Visible-light cameras depend on the scene giving them enough information to work with. That information changes with the time of day, weather and surroundings. Thermal imaging provides another view because it responds to infrared radiation rather than visible light, allowing temperature differences between vehicles, people and their surroundings to remain apparent when normal visual detail becomes harder to distinguish.
That does not make thermal imaging a replacement for conventional video. The two provide different information, and the usefulness comes from being able to draw on the right view for the situation.
On a clear afternoon, the visible image may tell an operator almost everything they need. Later that evening, or in difficult weather, thermal information can help preserve a clearer distinction between objects and the background.
For a highway that has to operate 24 hours a day, that consistency matters more than any single spectacular detection.
Almost Two Decades of Change
The A75 programme has gradually incorporated newer cameras, AI-based detection, redundancy between control centers, tunnel automation and closer integration with other operating systems. The current project material describes the relationship as a 19-plus-year partnership, built around continuous modernization and operational trust.
More recent work has included modernization using TrafiBot2, ITS Dual systems in the tunnels, and TrafiBot Dual AI cameras on highway sections. The stated aims are practical ones: improve system availability, reduce operational risk, improve detection and data quality, and give the road a platform that can continue to evolve.
That long-term view is perhaps more interesting than the idea of a suddenly “smart” highway. The A75 has become more intelligent gradually, as new requirements and technologies have appeared.
What Happens After an Alert
Detection is only useful if the information reaches somebody who can act on it.
The A75 operates with two control centers, and the system has been developed to support redundancy between them. Other functions have been added over time, including event-based video display, alerts and integration with tunnel SCADA systems.
That matters because an incident rarely exists neatly inside one piece of software. A stopped vehicle may require an operator to check the scene, communicate with another team, change signage or lighting, and coordinate the response on the road.
The camera is therefore only the beginning of the chain.
What matters to the person in the control room is that an unusual event is brought to their attention quickly, that they can see enough to understand it, and that the wider system helps them respond without having to piece the situation together from several disconnected sources.
There is still a person at the center of that process. The technology simply gives them a better chance of seeing the problem early.
Built Around Change
The A75 is not a controlled laboratory, which is precisely why it makes an interesting ITS case study.
It has tunnels and open highway, straight sections and bends, changing traffic levels and changing weather. The light at 8am in July bears little resemblance to the light at 8am in December. Yet operators still need a dependable view of what is happening.
After almost two decades, the lesson from the A75 is less about any one camera or piece of software than about designing a system that can adapt.
The road changes. The technology changes. What has remained constant is the requirement behind both: give the people responsible for the highway enough information to make a good decision, quickly.
