Drones, AI, 5G and Network APIs: a next-generation alert chain for airport security

Producing fewer alerts, but better qualified ones 

In the field, this new organization relies on unmanned autonomous vehicles (UAV) capable of generating a massive amount of information.

A drone can produce about 30 gigabytes of data per hour. Transmitting everything continuously is neither feasible nor useful», reminds Jad Rouhana. This is precisely where the network becomes strategic.

When the AI embedded in a drone detects an abnormal situation, it triggers a specific request to send a richer data stream, particularly in video, » explains Philippe Delbary, Head of Innovation Product at Orange. « The network then responds by guaranteeing the necessary bandwidth so that the operator can make the right decision.

The network thus becomes a full-fledged player in the alert chain: it ensures the necessary bandwidth to enable reliable verification. In an airport environment, where a confirmed alert can lead to the closure of an area for several hours,

the quality of the video stream is crucial both to avoid false alarms and to confirm real situations,” explains Philippe Delbary. For drones, this capability is fundamental to the very existence of the system. “Without prioritizing the data streams, the solution simply cannot function,” summarizes Jad Rouhana.

The relevance of an alert also depends on how readable it is.

An raw alarm is useless if it is not contextualized,” recalls Sébastien Roche, CEO of Optim.aize, a specialist in industrial and geospatial data. His team is developing a security digital twin, a true digital envelope of the airport, in wich existing data and information reported autonomous vectors are aggregated. “We contextualize the vectors in real-time: immediate environment,” he explains. “Within this digital twin, we correlate alarms, video streams and geospatial information in order to provide an actionable view to the command center.”

 The airport fully shares this requirement. 

The key is to avoid overwhelming people with incomprehensible alerts,” stresses Jérôme Morandière. Because for the airport, the quality of information takes precedence over quantity. “If a sensor sends an alarm every minute, the operator can no longer make decisions. What matters is the quality of the information, not its volume. 

When the network becomes a security tool

At the heart of this chain, the network also plays a security role. Concretely, the Orange API components make it possible to chain several security layers: authentication of the vector, geolocation, then activation of enhanced quality of service (or quality on demand).

When an alert is triggered, the operator starts by verifying the drone’s identity and location,” explains Philippe Delbary. “Only then is the video stream opened at maximum quality. We move from a passive network to a network that adapts to the business application,” he sums up. In other words, the application requests, and the network executes.

This double verification is a major issue:

We cross-check the information coming from the robot with that from Orange’s network to prevent any spoofing,” confirms Jad Rouhana.

Network-based geolocation provides an additional guarantee.

 A GPS can be spoofed, but a network coordinate cannot,” points out Philippe Delbary

Experimenting, securing, industrialize

Orange 5G Lab facilitated the launch of the project by providing 5G routers, and the Orange Livenet team deployed an experimental 5G network for testing. If the technology is promising, its large-scale deployment must the comply with the stringent constraints of the aviation sector.

Airports have extremely high cybersecurity requirements, and we will not be exempt from the framework of the European AI Act” recalls Jérôme Morandière.

For Sébastien Roche, the key lies in gradual rollout: “We need to validate the sensors, measure false alarm rates and test multi-vector orchestration in real-life conditions.”

This initial experiment also paves the way for other use cases:

It works for security in sensitive areas (ports, industrial, military or government sites, etc.) but also for industrial maintenance, for example to inspect Orange pylons. Today, these missions are still carried out by human operators, sometimes at the cost of serious accidents,” explains Philippe Delbary

In Bordeaux, the airport experiment sketches out a concrete trajectory for the future of security: drones and AI on the front line, humans refocused on decision-making, and 5G that no longer merely connects, but becomes a true tool for securing and managing alerts.