This interview was originally conducted in 2020, 3DR was since acquired by Kitty Hawk.
We’re very excited at Guinn Partners about the work Auterion has been doing in the commercial drone space. We would love to hear from you any thoughts on how Dronecode has evolved in recent years, and how companies like Auterion are building on top of what you’ve helped manage there.
Dronecode was created five years ago as part of the Linux Foundation to do what the Linux Foundation does best, which is to create industry consortia around open source. That means proper governance, legal structures, commercial-friendly licensing (“permissive”, such as BSD and Apache, not “copyleft”, such as GPL V3), and professional development practices such as public roadmaps. Once we established our focus, Dronecode grew around the PX4 ecosystem, including the Pixhawk autopilot hardware reference design, PX4 flight code, MAVlink, QGroundControl (QGC) and some related projects such as UAVCAN and NuttX.
Today, this ecosystem mirrors Linux itself. There is the code itself, which has matured at a dizzying rate thanks to one of the largest and most productive dev team in the industry, the companies that contribute to the code (both in funding Dronecode and assigning engineers to work on it), and the companies that have built a business around extending the open source PX4 code with commercial offerings. Foremost among them is Auterion, which was formed by the core team that created PX4 while at ETH Zurich. Auterion is like the Redhat of PX4, offering a commercial version of PX4 that goes beyond the open source core with cloud services, development tools and features such as obstacle avoidance and visual navigation.
Loads of companies have been built around this ecosystem, led by startups like 3DR, Auterion, Wingtra, Wingcopter, Modal.ai, Quantum Systems and others, which have raised nearly a quarter billion dollars between them. The platform has been embraced by giants such as GE Aviation, Intel and NXP. Recently, US concerns about Chinese technology have been a driving force and the US Army’s standardization on MAVlink and QCG in its small drone program is driving adoption by companies targeting the US government, such as Freefly, Impossible, and Vantage.
You’re working now with the FAA & ASTM on standards for drones. Can you tell us a little about that work?
In October, the FAA plans to phase out most of exemptions, waivers and IPPs that have been used to fly beyond Part 107 to date and instead encourage everyone operating outside of Part 107 to use a Type Certified (TC) aircraft, which is the most scalable way to approve such operations. To make that easier they’ve introduced the streamlined Durability and Reliability (D&R) TC process, which evaluates entire system performance, not components.
We (3DR) have one of the first aircraft (the H520-G) in that certification queue and we’ve been helping the FAA develop the process, ranging from the testing to the production certificate for the assembly factory. the way the FAA processes work is that the FAA says “what” it wants and then turns to an industry standard group (usually ASTM) to say “how” to achieve that. The FAA nominated me to lead the D&R standard at ASTM, and I’m pleased to say that it’s gone well and we’ll be entering balloting this month, with a goal of finishing before the FAA’s October deadline.
What is your ideal vision for the next generation of drone standards and regulations after this current generation expires?
The bad news is that it’s really been a mess to date. I’m frankly stunned that we still don’t have BVLOS and one-to-many drone operations routinely today in the US. The good news is that the insane logjam has reached a breaking point and as a result everyone, from Congress to the FAA, has accepted we need a different approach. Thus the performance-based D&R certification…
3DR made a niche for itself in recent years working with the Department of the Interior and releasing a drone with Yuneec that is fully assembled in the US. Do you think that restrictions on the use of Chinese drones in the US is something that’s here to stay? If not, how do you feel American companies can best compete with DJI?
Sadly, I do think that paranoia about China is here to stay. This predates the Trump Administration and it will postdate it. Although China is indeed a strategic competitor, I’ve seen no evidence that DJI (or any other Chinese drone company) have so far been drawn into that fight. That said, I’ve lived in China and I know that things could change with one phone call from Beijing. So like it or not, that’s the world we now live in.
There’s a lot of discussion in the White House, the Pentagon and Congress about which elements of a drone are security-critical and we, along with many others, have had our say. At the end of the day, I think it will boil down to software and some critical hardware pieces, such as radios and perhaps flight controllers that should be assembled in the US to ensure that there are no “backdoors”. So be it — we, like many others, have the US factories ready to comply with that.
Looking back on your time at 3DR and with Dronecode, what are some recommendations you have for new entrepreneurs in the drone space as they build their companies?
It’s very tempting to be another “systems integrator”, putting together other company’s components into yet another bespoke solution for some imagined need. Although that may be fine for the onesies and twosies, it’s no way to build a scalable company. My advice is to pick a vertical (from survey to delivery) and spend a year *really* understanding it — what’s broken, what’s fixable, what the underlying unit economics are. Only then start working on a drone solution. We’re now at the stage when the technology should mold to the marketplace needs, not the other way around.
3DR’s Site Scan is now a part of ESRI, how did that acquisition come about?
We started several years ago recognizing that GIS was distinct from AEC and that ESRI’s platform (ArcGIS SDK) was increasingly easy to develop for. The first integration efforts went well and we became more aligned with the ESRI team and their long-term strategy. Then the US government concern about DJI kicked in and we realized that as a US company the ability to serve government (from Federal to local) was going to be a competitive advantage. This shifted our focus even more to ESRI, which is primarily focused on government markets. The rest was simply a matter of even more aligning teams, technology and strategy. Fortunately, they were thinking the same way. The planets aligned…
What recommendations do you have for other aerial data and drone startups who are looking to eventually be acquired?
Understand your aquirerer! Companies are bought, not sold.
Online, it’s become obvious your passion for autonomous drones has lead into a passion for DIY autonomous vehicles. What are some resources interested people can use to learn more about that technology and community?
More than ten years after I launched DIY Drones to help kickstart a bottoms-up reinvention of the aerospace industry along “cheap, easy and open” principles, I think we can declare victory and move one. Although drones are largely a solved problem from a technical perspective, there are plenty of other autonomous systems that are not. Ironically, although cars are just 2D, compared to drones in 3D, they are harder: you can’t count on GPS guidance and the navigation space is crowded, with little tolerance for a few meters of navigation error. That requires computer vision and AI, which is much harder than the relatively conventional Bayesian control theory we used with drones.
So it’s time to move to next level of challenge! I now spend most of my spare time on running an autonomous car racing community, which applies the magic letters “DIY” to self-driving cars. That means the same thing: cheap, easy and open; under $400, doable in a weekend, open source.
Here are some resources to get started:
- DIYRobocars website, with links to groups around the world
- The DonkeyCar project, our main reference platform
- Meetups (virtual races in simulators and real races in the Bay Area)