Guilty Until Proven Innocent
Updated: Oct 23, 2021
We had an unique opportunity to conduct an interview with the CEO of AFuzion Inc., Mr. Vance HILDERMAN, and learned about the future of certification in the UAV market. We are thrilled to have had this interview with him. We are pretty sure that you will also learn a lot from the following interview notes. Please contact us at email@example.com for any questions and requests.
EV: What are the Top Three emerging trends for Drones and UAV’s you’ve seen lately?
VH: There has been tremendous growth in commercial drones/UAV’s recently. While consumer-based (less expensive and shorter range) have been gaining popularity the past decade, the commercial world has seen unprecedented recent growth. Top trends are electrification, LIDAR-enhanced navigation, commercial and military cargo deliveries, commercialized avionics subsystems off-the-shelf including software and hardware for multiple platforms, and smart “wingman” for military support including AI.
EV: What is your company’s relationship to UAV’s and drones?
VH: About 30% of AFuzion’s work is with UAV’s and eVTOL; we provide engineering development and certification services. Drone/UAV certification can be both expensive and challenging to understand so we have a freely available technical whitepaper which is quite popular and can be downloaded online.
UAV certification primarily addresses public safety, so the more a particular UAV or drone can negatively affect that safety, the more rigorous will be its certification. So small consumer oriented craft weighing 1 or 2 Kg might have to be registered in some jurisdictions, but not certified; instead they have to follow common sense rules about staying away from the public. However, larger UAV’s, or flying beyond line-of-sight, or flying near the public or other aircraft requires increasing degrees of certification. “Certification” primarily ensures product reliability, whereas “Operation” primarily addresses rules for flying the UAV. Today each country has different rules for both Certification and Operation however we’re seeing a merging of those rules. Traditionally the FAA has led the world on aircraft certification but today Europe, Australia, New Zealand and many other countries are increasingly active in setting their own rules and also contributing to the emerging international standards.
EV: Are the emerging UAM/eVTOL aircraft related to drones and UAV’s?
VH: Eventually, perhaps 10-15 years from now, we’ll have autonomous UAM aircraft; technically we can do that today but the reliability and operational rules are still too immature to allow autonomous (no onboard pilot) aircraft. However, autonomous small cargo aircraft are operational today for special purpose missions: those will lead the technological trend since there are no humans onboard so “flight safety” can be significantly more simplified. But yes, there is a relationship between unmanned and UAM aircraft: UAM must (let me emphasize that, MUST) first be human-piloted prior to a downstream transition to unmanned UAM. Think about it: technically we can fly commercial aircraft without a pilot; avionics has had those capabilities for over a decade – it’s almost “old”. But why don’t we? Or why don’t we even reduce the standard two-pilot passenger aircraft to a single pilot? Someday we will. And someday we’ll have 30-day development times for new pandemic viruses along with cold fusion. That is simply not the aviation certification environment of today – far from it. Experimental aircraft take fast and bold steps; commercial passenger aircraft evolve MUCH more cautiously. Now, we’re already demonstrating autonomous cargo and passenger aircraft capabilities and they work, truly. We’re proving them today in remote locations and low population density areas. Hint: we’re doing that in countries with less litigation culture than the USA and less-rigid certification culture than the EU. So first we’ll complete certification of cargo drones in low-litigation countries over unpopulated areas. Then we’ll gradually progress to more population, then passengers, then to USA / Europe. So if you’re a passenger awaiting an unpiloted and certified aircraft in the USA or Europe, start signing your children up: they’ll be the first paying passengers … 😊
EV: Can you summarize the Costs versus Benefits of following these new UAV/drone development and certification rules?
VH: New aircraft certification costs are huge. Let me rephrase that: NEW AIRCRAFT CERTIFICATION COSTS ARE HUGE. Consider companies developing simple software like Microsoft or Facebook ( the reader should be smiling now, as they know Microsoft and Facebook software encompasses hundreds of millions of total SLOC’s; if you don’t know what a SLOC is, then you’re not a software engineer so it frankly doesn’t matter). But aviation software is not Microsoft Flight Simulator: almost every SLOC onboard an aircraft has detailed requirements, design, and tests replete with evidence. Remember, in most countries the legal system states “Innocent until proven guilty” (in theory, if not in practice). But in aviation, it’s the opposite:
“Guilty until Proven Innocent” the aviation engineers at AFuzion say. So be it.
Now, aviation certification regulations produce many benefits for aircraft and systems developers: proven high quality, proven quantitative reliability/MTBF, ability to perform Root Cause Analysis, greater reusability, faster integration time, and more scalable solutions. On the flip side, all of these add significantly greater up-front costs: if someone is a hacker, they’ll find certification increases their costs by 1000%. But if someone has good development practices with actual basic planning, requirements, design documentation, thorough testing, and independent reviews, then added certification costs are much less, perhaps an additional 30-50%, since the majority of certification tasks were thus already handled. If someone wanted details, our new whitepaper on UAV certification costs/benefits could be helpful – free download here.
EV: How does AI fit into the UAV/drone sector?
VH: Artificial intelligence is both a huge opportunity, but also a huge unknown for UAV/drone certification. For passenger aircraft, we currently must avoid true AI since it is outside the scope of today’s “provable determinism” requirement for passenger aircraft certification per CFR 14; but we’re progressing on that so should have answers and operational flight software in a few years. But drones are somewhat less regulated and also since lacking an onboard pilot can more greatly benefit from AI. So we’re seeing AI already in simulation, mission planning, and even onboard “loyal wingman” type applications. I can’t go into greater details due to NDA’s, but AFuzion is supporting several aircraft projects utilizing AI and it shows huge promise. The biggest challenge to AI is provable safe deployment and there are several approaches to ensure this; the article here has greater detail: https://www.coderskitchen.com/artificial-intelligence-in-safety-critical/
EV: Can we use AI for flight control and navigation of commercial passenger aircraft including pilotless?
VH: Great question and the easy answer is “Not Today … Yet.”! Both Airbus and Boeing are devoting great resources to it, and many worldwide governments are also involved. There is even a joint international commercial aviation AI consortium pooling resources to help bring AI to onboard aircraft systems. In an over-simplified view, one could say Boeing builds Porsches which are driver’s cars, e.g. the pilot really controls the aircraft. In contrast, Airbus builds Cadillacs where the driver is insulated from the road and the avionics fly the aircraft. Vastly over-simplified, but some truths. If you have great pilots, you’ll prefer Boeing. If you have weak or average pilots, you may prefer Airbus. Funny thing about “average” … on average, the average pilot is … AVERAGE. The challenge with highly automated aircraft is a pilot with 5,000 hours really only has 500 hours of flying the aircraft; about equivalent to an American 19-year old inexperience car driver. Aircraft are becoming increasingly more automated, even Boeing’s, and that increased automation causes challenges (737MAX). So we need better systems engineering via ARP4754A and ARP4761, better simulation, and better pilot training. Everyone must recognize this. But therein lies the beauty of AI to help mitigate the obvious evolutionary challenge of increased automation/complexity. I predict we are 10-15 years away from seeing “true” AI (not your parent’s partial Artificial semi-intelligence) in the cockpit flight control and navigation systems.
EV: You have a new book out, available on Amazon, called “Aviation Development Ecosystem”; how does that relate to UAV’s and drones?
VH: Yes, since aviation’s fastest growing segment is UAV’s and drones, the 30 aviation engineers who contributed to or reviewed the book all wanted to ensure coverage. Designing a truck has many similarities with automotive design: 90% the same, 10% different. Same for UAV’s and piloted passenger aircraft: 90% the same, 10% different. My latest book was Amazon’s aviation best-seller for its first month – Amazon doesn’t provide buyer feedback but we received about 200 emails from the 3,000 buyers that first month and we noticed that half of those book buyers were working on UAV / drone projects. All the comments were gratifying (except one reader hated the book; he wrote his review exactly 20 minutes after the book was first released digitally so perhaps he didn’t actually take time to read the book or he’s an amazing speed reader to have digested 450 pages in 10 minutes then 10 minutes to write a review). So the book seems to be well-used out in the UAV/drone sector.
EV: A company has a CMMI Level-3, basic training of DO-178C, ARP-4754A, ARP-4761 etc. This company wants UAV&UAS system development. Does this company need additional special training?
VH: Great question: CMMI Level-3 compliance is the single best initiator of good system/software development practices and frankly EVERY company should follow, even if they’re making cappuccino machines (which is considered a life-critical device in my house). But seriously, a CMMI Level-3 rating should reduce the required training needed for aviation development certification success by 40-50%. But developers will still need additional aviation development certification training in DO-178C, DO-254, ARP-4754A, and ARP-4761 since 50% of that is outside the scope of CMMI. The Society of Aerospace Engineers (SAE) provides training in all these topics multiple times per year so that is a good place to start (they’ve selected solely AFuzion to provide all their training so we provide training of about 3,000 persons per year in that).
EV: Is there any regulation/certification for the swarm of UAVs?
VH: Wow, another great question. A good answer would be “Yes, and we’re publishing that information soon”, but alas: that would be false. The real answer is “That is a great question as it addresses a great need for increased guidance of swarms. However we first need to finalize various international single-UAV/Drone certification guidance before moving on to swarms.” So, I’m very sorry to say, “we’re not there yet.” Hint: look to Australia for more near-term word on this topic (Australia, for the less-experienced, is a hot-bed of innovation and leadership in aviation, safety, and new technologies. Very forward thinking and Afuzion has multiple ongoing projects in Australia, both civil and governmental. Very advanced.)
EV: Is there any regulation/certification of UAVs used in agricultural medicate?
VH: The agricultural area has largely been hands-off for drone certification, unless you happen to be farming next to an international airport. But farms are general in more remote, less-air-travel regions and flying very low-level. There simply are not many farms near Class A, B, C, or D airspace obviously, so agricultural has been low-visibility, low concern. Now, medical is a different story: medical can often be safety-critical, ferrying victims, patients, organs, or urgent supplies to varying locations. We’re seeing major eVTOL and UAV capabilities exclusively devoted to this medical aviation sector and so far there are not ample special rules for medical related aircraft: they simply must follow all the regular associated (non-medical) aircraft certification and operational/flight requirements.
ABOUT VANCE HILDERMAN (CEO, AFuzion Inc.)
BSEE, MSEE, MBA
Founder of two of the world's largest avionics development services companies
Developer of the world's first training in DO-178 and trainer of over 8,000 engineers in 45 countries in DO-178, DO-254, DO-278, and DO-200A
Primary author of the world's first, and best selling, book on DO-178 and DO-254 (available at bookstores worldwide)