Wednesday, June 23, 2021

How AI and connectivity could unlock a new drone age for the industry

The ability to fly connected drones “beyond visual line of sight” on automated missions could transform many business operations, experts believe.

Today, if you want to inspect a building site or other installation with a drone you need a pilot standing nearby to keep an eye on the craft at all times and control it over a radio connection.

But, according to John McKenna, chief executive of sees.ai, a tech company working on how drones can be operated remotely using artificial intelligence, this means that a pilot “has to drive to the site, go through a health and safety induction before flying, and then after flying, pack up and drive back to the office.”

This “wastes a lot of time”, he says, and is preventing drones from fulfilling their potential as useful monitors inspecting places that are difficult or dangerous for humans to reach.

So his company is working on a project with Tokyo-based TerraDrone, and with Vodafone as technical partner, to see how drones could be operated remotely “beyond visual line of sight” (BVLOS) without compromising safety.

Remote benefits

The benefits of remote drone operation are obvious.

Falls from height are still the main cause of fatal accidents in the workplace, so sending drones out to do dangerous inspections rather than people could save lives.

Drones can monitor buildings, power lines, wind turbines, disaster zones and difficult-to-reach places much more quickly and efficiently, and thereby help experts resolve problems faster. They can also deliver much-needed supplies, such as blood or COVID-19 vaccines, to remote areas in a fraction of the time it would take conventional transport.

And if they can do all this while being operated remotely from a centralised control centre rather than by pilots in the vicinity, it makes things even more efficient.

As it stands the UK’s Civil Aviation Authority (CAA) only allows remote BVLOS drone flights in restricted cases, but it is working closely with industry to come up with a solution that is safe and reliable.

Headwinds

But piloting drones remotely in industrial or dangerous environments presents challenges.

“If you want to remotely operate a drone in and around a construction site or an oil rig in the North Sea you need to have very precise control of the system,” explains Mr McKenna.

Drones can lose network and GPS signals going in and around metal structures, and in emergencies first responder drone users can face network congestion, making controlling the drone difficult.

So sees.ai is developing computer vision technology that will enable a drone to fly independently using LiDAR (Light Detection And Ranging), cameras, and inertial navigation sensors in real-time, even when communications are down, so that it can navigate without colliding with unmapped objects.

The company has lined up early tests with construction company Skanska; with Sellafield in Cumbria; Skanska Costain STRABAG, the joint venture building Phase 1 of the HS2 railway line between London and Birmingham; and Atkins, an engineering contractor working near live public roads.

Command and control

There is “a growing consensus” that mobile network operators have a key role to play unlocking the commercial drone industry, says Jonathan Reid, Vodafone UK’s Defence Sector Head and UK Drone Lead.

Vodafone’s Radio Positioning System lets drones use signals from network masts to provide a back-up navigation source, says Mr Reid. And if drones have SIM cards and modems, it’s easier for regulators and law enforcement to identify their owners.

For now, sees.ai is working to unlock use of 4G and LTE networks for command and control communications, but they’re also working out how 5G will increase drone capabilities.

5G will provide “much reduced latency, and increased bandwidth”, says Mr McKenna. On the downside, very-high bandwidth 5G can degrade more with distance and rain. 5G is “still evolving, especially for the Internet of Things” (IoT), he says.

“Having a partnership with Vodafone means we can tease apart the constraints and opportunities better than we could ever do if we were just an outsider trying to figure it out,” he says.

Engaging in projects that evolve and test the benefits of 5G are key to unlocking the technology’s potential, says Mr Reid.  As an example, he cites the 5G maritime test bed that Vodafone is delivering for Smart Sound Plymouth.

As well as a technical partner, Vodafone was sees.ai’s first-ever customer. The company designed a system that “flew autonomously around a mobile mast, and could map the radio spectrum coming out of the tower” in three dimensions, says Mr McKenna. Air traffic navigation company NATs is a technical partner, too.

Flying ahead

Sees.ai is “within half a year” of flying test missions on industrial sites, such as HS2 and Sellafield, and with emergency services such as the police and Lancashire fire and rescue, says Mr McKenna. The tests will challenge the technology to cope with harder and harder conditions and terrains.

In November 2021, the company and seven partners received a share of the Future Flight Challenge – £30m in grants from government agency Innovate UK directed toward promising new aviation projects.

Technology enabling drones to operate safely beyond the visual line of sight “is the key technical enabler” to unleashing industrial drones, says Gary Cutts, the Challenge’s Director.

Once drones can sense and avoid obstacles automatically to a high enough standard for aviation authorities, they’ll be integrated with fully autonomous systems, suggests Jonathan Reid. A drone could fly autonomously on a pre-defined route and schedule, for example, and its AI could detect and identify a tree that has fallen across a power line. This would then trigger an automated alert to a removal team.

AI, central control and high-speed connectivity will unlock a new drone age for industry.

Follow @VodafoneUKNews on Twitter.