Intersections need safety layer that aviation built decades ago

By Georges Aoude

Derq's Georges Aoude suggests the road transportation industry can learn by looking to the skies

Intersections remain one of the most predictable points of risk in our transportation networks. Road safety has long focused on better vehicles, stronger enforcement, clearer signage, and more public awareness. All of that matters. But it still leaves too much responsibility on individual road users to detect conflict, interpret intent, and react in time.

Aviation took a different path. Aircraft don’t operate in isolation. Safety in the air depends on structured communication, shared awareness, and infrastructure that identifies and resolves risks before they become disasters. Pilots are highly trained, but aviation safety wasn’t built by assuming perfect human judgment in every moment. It was built by creating a system that reduces uncertainty and supports better real-time decisions. Aviation relies on mandatory communication and shared system design.

Roads largely don’t, so that gap is especially clear at intersections, which concentrate the conditions that make road safety difficult: competing movements, limited visibility, changing signals, pedestrian crossings, cyclist interactions, and unpredictability. These are everyday scenarios, not edge cases.

Conflict and understanding

Today, most traffic infrastructure can control movement, but it can’t truly understand it. It may know the signal state, but not whether a vehicle is unlikely to stop, whether a pedestrian is entering conflict, or whether a near-miss pattern is emerging that signals deeper risk. In most cases, agencies learn where the danger is only after crashes accumulate.

“Safer systems don’t depend on isolated actors making
perfect decisions with incomplete information”

In aerospace, the goal isn’t simply to respond once something goes wrong. It’s to maintain a shared, real-time picture of the environment so conflicts can be detected early and addressed before they escalate. Aircraft continuously share position and movement, and ground systems help create a common operating picture. That shared awareness helps humans and automated systems spot and resolve conflicts before paths intersect. Roads need their own version of that shared safety layer.

That doesn’t mean treating streets exactly like airspace. Roads are messier, more open, and more variable. But the principle still applies: safer systems don’t depend on isolated actors making perfect decisions with incomplete information.

Risk and movement

At intersections, intelligence needs to live in the infrastructure itself. An intelligent intersection can detect road users in real time, interpret movement and identify elevated-risk situations before a crash occurs. Sensors and edge AI can help infrastructure move beyond passive control—seeing conflict patterns, detecting dangerous approaches, and enabling earlier warnings or interventions. Detection, prediction, alert, and action can work as a continuous loop, supported by AI-powered ITS and V2X, where relevant.

This matters because the next step in road safety won’t come from asking more of road users. It’ll come from building environments that provide greater visibility, richer context, and earlier warning signs of danger.

The tools already exist. What’s still missing is the willingness to treat intelligent intersections as infrastructure rather than pilots or operational add-ons.

If aviation teaches us anything, it’s that safety improves when communication is structured, awareness is shared, and the system itself is designed to anticipate conflict. Roads, especially at intersections, still ask too much of individual perception alone.

We should keep improving vehicles. We should keep enforcing laws and educating road users. But if we want a meaningful step change in safety, we need infrastructure that does more than manage movement. We need infrastructure that helps prevent harm.

Intersections don’t have to remain places where visibility is partial, and prevention comes too late. We can make them more intelligent, more responsive, and far safer—today.