American Robotics CEO Reese Mozer has no beef with drone deliveries, but he thinks all the hoopla surrounding aerial transport of burgers and burritos is drowning out news about farther-reaching UAV activities that are dramatically changing the way businesses operate. He tells DroneDJ about that transformative innovation, and how American Robotics’s (AR) leading role in the complete automation of critical drone services to industry is set to take wing.
According to McKinsey, 660,000 commercial drone deliveries were made over the past three years – nearly 500,000 of which were made in 2021 alone – with those numbers set to jump to 1.5 million this year. While that’s all fine and good, Mozer notes much activity in that global estimate is taking place outside North America, particularly through multiple medical supply flights that UAV logistics companies like Zipline and Swoop Aero operate in African nations each day. Commercial drone deliveries, he says, remain limited in scale.
Nevertheless, Mozer adds, that action manages to generate sufficient media and public excitement to divert attention from the more complex, vital, and – in total financial terms – valuable surveying and inspection services drone automation provides heavy industry, energy, railroad, and infrastructure operators. And that’s precisely the UAV sector activity he predicts will begin taking off and turning heads this year.
Software-based wireless broadband company Ondas Holdings agrees – which is why it acquired AR last year in a deal valued at $70.6 million. Just this week, it added artificial intelligence (AI) railroad analytics software company Ardenna to the growing team. The objective in that is clear: use AR’s AI-enhanced UAV-in-a-box Scout platform and analytics software to provide a full range, surveying and inspection system for industrial clients with complete drone automation.
AR does that, moreover, using the only Federal Aviation Administration (FAA) waiver granted to date for regular beyond-visual-line-of-sight drone flights without a human ground controller or visual observer.
Mozer tells DroneDJ why all that’s a bigger deal than getting coffee and cake from the sky.
Reese Mozer: Drone deliveries are getting too much attention and have less potential than marketing would have you believe. I can understand the public appeal of getting a burrito delivered to you by drone, but when you actually drill down into the details about what it takes to do that kind of operation versus others, you start to see that data collection drones have a bigger place in the near term, if not in the long term as well.
The commercial drone segment can be divided into different, high-level groups: drone delivery and data collection. We play in the latter of those two, and the number of markets and use-cases in that is pretty expansive.
This category of drones is about digitizing the physical world for all sorts of industrial applications: agriculture, mining, oil and gas, rail, solar, nuclear power plants – basically any industrial activity with some sort of physical asset presence. This is about the ability to inspect and monitor those assets at a very high resolution and a pretty high frequency to be able to detect methane emissions or rail defects that might cause a train derailment. So the value is arguably a lot higher, and I think we’re going to see a much quicker proliferation of autonomous drones in the inspection category than of those in the delivery category.
If we can expect a bigger boom in non-delivery services, which aspect will be most valuable to industrial clients: the aerial capabilities of drone operation, or the automation of those?
Mozer: Automation is at the core of everything we do, and is really at the core of our pitch and value proposition. When you look at all the enterprise use-cases for inspection, we estimate 90% or more of them require full automation to makes sense.
Imaging of an asset once or twice a year has little to no value because you’re trying to detect problems not knowing when they’ll occur. It’s about inspecting how far apart the railroad tracks are, are there any broken ties, any missing nails – anything that might affect the safety of the track. In gas, it’s looking for leaks, methane emissions, corrosion on pipes. On solar farms it’s looking for broken cells, which often come up in the form of a hot spot.
To do that you need to cover a lot of area, very frequently, and at a high enough resolution for AI to make sense of it. That means you basically have to run these drones on a constant basis every single day. Doing that with human labor, if you scale that across every asset, it quickly doesn’t make sense.
There over 900,000 well pads in America, over 500,000 miles of pipeline, 140,000 miles of track – eventually every single inch of those assets are going to need to be surveyed, every day. And doing so while having a pilot on the ground with a remote control, watching the drone constantly just doesn’t make any sense. So automation is the core that unlocks scalability in most use-cases.
How close is 100%, 24/7 automated drone services from going big time?
I already think it’s ready from a need and customer perspective. But that hasn’t been possible until our system matured – both in terms of technology and regulatory approval. For the past 10 years this whole space has been stuck in this trial period where all companies can do is invest in a handful of drones and start to learn how these things might help them, waiting for automation to mature and for that technology to be approved by the FAA.
That’s the position we’re in now, so I think we’re finally at that inflection point. The next five, ten years will look a lot different from a user perspective than the last five or ten. We’re still very much on the ground floor, but at an inflection point – and 2022 will mark that in the history books for the commercial drone industry. You’re going to start hearing more about actual autonomous operations with industrial customers this year, and that will start to grow exponentially next year.
Why is it American Robotics still the only company with an FAA waiver for automated BVLOS flight?
This technology is really hard, and the FAA is really hard, so it’s both things on top of each other. I also think there’s an under-estimation of just how difficult it is to actually create a system that safely and reliably live in the field, on its own, and operates 10 times a day, every day. That’s an incredibly challenging problem for all sorts of hardware and software reasons that only an incredibly small number of companies in the world can even hope to overcome.
The other aspect of the answer is there are a number of proprietary technologies we have developed, some of which are not yet public, that directly allowed us to get this waiver. And we had to get approval not only for flight beyond visual line of sight, but to have no visual observers on the ground, ever – during, after, or before for the 100-year FAA default rule of human pre-flight check. So that’s another aspect of the complete drone automation process we had to overcome.
What has changed most since your acquisition by Ondas – use of their private communications networks rather than legacy providers?
That’s certainly one aspect of it, and the value of that will increase over time as our systems scale in greater numbers across bigger areas that require those kinds of longer-range communications on licensed spectrum. Another aspect is that we’re a public company, with access to significantly more resources than before, and our team has grown 500% over the past year as a result. So that allows us to prepare the company to start delivering these systems at scale.
Also, one of the reasons we chose each other was a shared theme of industrial data – collecting, analyzing, and transferring data – and how can we work together to improve our solutions even more for all our customers, many of which overlap. Ondas works in the rail space, which is one we’re interested in. They have interests in oil and gas, which is an area work in. We share similar activities aiming for the same objective: a new era of industrial data solutions.