We were promised drone deliveries. What's the hold-up?

We might have to wait a bit longer before we start getting our online purchases dropped off by drones. An expert explains.

Dr Yaron Kanza, Ph.D., a computer science researcher at AT&T Labs-Research is, along with his colleagues Dr. Tamraparni Dasu and Dr. Divesh Srivastava, an author of the 2018 award-winning research paper called "Geofences in the Sky: Herding Drones with Blockchains and 5G". Dr. Kanza was kind enough to tell us about the aerial infrastructure that needs to be put in place before the upcoming mass adoption of unmanned aerial vehicles (drones) by e-commerce businesses, media companies and the general public becomes possible.

Dr. Kanza, we are seeing more and more drones in the sky these days, both in terms of private and commercial usage. Where are the Amazon drone deliveries?

Indeed, there are more than 700,000 drones already registered in the United States and the numbers are rapidly increasing. Very soon we might see millions of drones completing all sorts of tasks. However, the regulations we have today state that drones should always be operated within line of sight, which prevents many of their most useful applications.

Such as drone deliveries?

Exactly. You can't really send a drone to deliver a package across town, for example. It might endanger people, damage property or cause an accident. In order to be able to use drones for more serious tasks like that, we need traffic control rules and regulations that deal with the various ways drones can become a threat to public safety.

You mean something like air traffic police?

It might be tempting to compare drone traffic with car traffic, in the sense that many cars on the road are able to operate without constantly crashing into each other. Unfortunately, we can't use the same type of regulations for drones. When there's a heavy traffic jam, for example, cars have no problem standing in one spot for a while. Drones, however, would not be able to hover indefinitely in the air. At some point they will simply fall out of the sky right on people's heads. Without proper regulation, the airspace might get too crowded and drones can start crashing into each other. But it's not just about regulation – when a drone is on a mission it should be able to complete it without colliding with other drones or aircraft. This requires a lot of meticulous planning.

Then shouldn't we have something like computerized air traffic control towers that deal exclusively with drone flight paths?

We can work with this comparison too. Think about airplanes today – if they want to fly from one point to another, they need to submit a detailed flight plan to the FAA or some other organization that can authorize their flight path and only then can the flight be carried out. This is a tried-and-tested way of managing airplanes. However, it does not scale and does not provide sufficient flexibility. On the other hand, we need to prevent drones from crashing into each other, veering off course, or flying into crowded areas. We need to do all these things simultaneously. 

Dr Kanza believes, "We need a new way of thinking about how to control drone flights on a wide scale."

Don't we need some kind of an airspace management agency coordinating all those drones that fly beyond line of sight?

Yes, but drone management doesn't have to be centralized to be efficient. In fact, what we at the Data Science & AI department of AT&T Labs Research are suggesting is to use blockchain technology in combination with 5G, in order to be able to accurately manage drone traffic in a truly decentralized way. Each drone will be able to reserve its own airspace corridor before taking off, which will guarantee a safe flight. As we know from learning about cryptocurrencies, blockchains were designed to prevent double spending. That's what actually makes cryptocurrencies useful as a means for value exchange – you can't use the same money twice. We can leverage this unique feature of blockchains to allocate flight corridors to drones, guaranteeing that no other drone will be able to fly through a corridor that has already been reserved by another drone. Now imagine millions of these airspace corridors, computer-operated in a decentralized way, allowing for millions of drones to follow their own flight paths without crashing into each other. All of this is possible today through a combination of blockchain technology, geofencing and 5G.

Can you elaborate on the role of geofencing?

The main idea is to partition the entire airspace in a decentralized way via blockchain, but there must also be certain areas that have to be controlled by a centralized entity or organization. These areas will be defined by geofences. Such geofenced air corridors can be used in emergencies only. For example, if there's a need for emergency helicopters or emergency drones to evacuate injured people, they can make use of these corridors because there won't be any danger of a random, unauthorized drone flying there. Geofencing can be utilized not only to define these “emergency lanes" in the sky, but also to outline the borders of no-fly zones. These will be areas that drones should not fly over, such as airports or government buildings. In residential areas, geofencing can be used to limit the time when drones can fly. Due to the noise or safety concerns, drones can be permitted to fly over residential areas only within certain hours of the day. Geofencing are also useful for protecting tall buildings from stray drones crashing into them. Also, if there are expected bird migrations in certain areas, temporarily geofencing them off makes a lot of sense.

You also mentioned 5G?

Yes. In places where there are no obstacles, such as more rural areas, we wouldn't really need 5G, because GPS-powered geofencing works fine. In dense urban areas, however, there are usually buildings that could block GPS signals. We should be aiming for a much more accurate geofencing granularity. With 5G, we would have antennas that serve as beacons providing location signals in a more accurate way. In the future, we will have a very reliable 5G network with much lower latency and much higher bandwidth compared to today. In addition, 5G could support the networking needs for communication with the drones and interactive responses to unexpected events. We will leverage this 5G network to augment our geofencing. This way, we will provide all drones with the most accurate and granular location awareness required for complex navigation, especially over dense urban areas with heavy air traffic.

Sounds like you have the technical solution down. Once we build it, can we finally start getting drone deliveries? Please say yes.

Technically yes. However, we will still need a regulatory framework to be put in place by the government. Hopefully it won't take quite as long as building the technical part of the solution, but who knows?