NASA Autonomous Operations Workshop

Former Acting Deputy Administrator, Carl E. Burleson (January 07, 2018–August 12, 2019)

Thank you for that wonderful introduction, PK. Good morning everyone. It’s nice to be in this beautiful part of California where two of America’s finest products – advanced automation and wine – are crafted. Today’s agenda is an engineering deep dive on the former, which for this economist, might mean a glass of the latter when it’s all said and done.

The focus of this workshop – Enabling Autonomous Operations – is of deep interest to us at the FAA. It is key to the evolution of the National Airspace System. How we integrate autonomy and new technology in general will fundamentally reshape how we manage, regulate and oversee a variety of new forms of transportation and add to the capability of traditional users.

And make no mistake, highly autonomous operations and other sweeping transportation innovations are coming – it’s inevitable. Six months ago, who would have thought we’d see three rocket boosters simultaneously stick their landings back on earth after a Falcon Heavy launch?

It’s happening. The question for FAA and other regulators is how do we introduce groundbreaking technologies into the safest form of transportation on the planet, wisely, safely and in a reasonable timeframe? And I’m glad you’re having this meeting as the work you all are doing will help inform us how we tackle this challenge.

The pros and cons of this future remind me of a conversation I had with my sons on a stormy evening long ago. 

It was a sticky August night in DC (we sometimes get those in DC…) and a raging thunderstorm had knocked the power out. I took the boys out to our screened-in porch to watch the driving rain, the wind and the lightning show.  My adventurous 12-year old pleaded, “Hey Dad, can I go out and play in this?” My younger son was sitting very quietly in my lap for the longest time.  I finally asked him- what he thought about the storm. He says, “Dad, is this any way for a seven-year-old to die?” That’s reality–two boys, from the same genetic stock, having two polar opposite views of the same phenomena they were observing.

Their divergent views of the world is a perfect analogy for the challenge we in the aerospace industry face with the rapid development of a set of new technologies that are paving the way to increasingly autonomous operations. Some see an incredible opportunity for development and new markets and services.  Others see this as undermining safety and adding new levels of risk into the system. As a regulator, that’s what we have to balance. How do we find the middle ground between those two points of view?

NASA is no stranger to the benefits, and often the necessity, of autonomy. The satellites you built and launched starting in the late 1950s had to be largely autonomous for the obvious reasons. As the onboard analog systems were replaced by digital microprocessors, miniaturization and health monitoring capability took a giant leap. Engineers were soon able to build onboard error detection and correction systems that required no intervention from ground controllers.

For manned aviation, autonomy kicked into high gear in the late 1960s with the certification of the first autoland systems. Later the flight engineer was replaced by computers, and when you look at new models like the Boeing 787 or Airbus A350, you’ll see an aircraft that is essentially a computer with wings. It will go farther still.

How fast we get there is the question. In commercial aviation, no new technology premiers before its time. The quantum leaps in capability or safety, for example fly-by-wire, TCAS or ADS-B, typically take two decades or more between first research and operational readiness. That’s also true of the ground automation systems that keep 50,000 flights a day running at peak performance 24/7/365 in all manner of weather.

I’m not sure we have the luxury of so much time any more.

Silicon Valley – right up the street – has introduced us to new technologies that are rapidly changing aviation and transportation whether we like it or not. If your cell phone is more than a year old, is it still cool? Probably not. And it’s a good bet the most important accessory in your private aircraft is now an iPad or tablet that does most of the navigating and other housekeeping.

The innovations are not going to slow down and that’s a challenge for the FAA. If we’re going to keep aviation on the cutting edge, delivering the most it can for the economy, we’re going to have to figure out how to integrate these new technologies, all with different underlying paradigms, and figure out the acceptable failure rates in our ten to the minus ninth world.

When the unmanned aircraft revolution started, we weren’t ready, and it was a hard lesson. 

The fact is, UAS caught us flat footed. There were some people in the FAA who saw it coming, but no one saw it coming so fast and so massive in numbers. We have been digging out ever since. As of last week, we have registered almost 1.4 million drones, nearly 400,000 of which are for commercial operations. A good many of these are small, semi-autonomous systems that you fly with your finger and a smart phone. UAS breakthroughs will see their way into urban air mobility vehicles. I’m sure we’ll see some of those innovations as part of the NASA Grand Challenges that start next year.

Automation in cars and trucks is in its early stages, about where we were with the first autoland systems in the late 1960s, but catching up with immense speed. By some estimates,  semi- or fully-autonomous ground vehicles – cars, taxis, buses, trucks – will represent a market potential of nearly half a trillion dollars in seven years. That’s up from about $50 billion this year. The DOT is preparing for this evolution by working with the industry to incorporate a slate of best automation practices.

And what happens when automation and computers enable those autonomous road vehicles to take to the air. It’s a crossover that will raise a wide set of thorny issues for manufacturers, drivers and regulators. I suspect that what’s allowed in terms of fatalities on the land today will be unthinkable for urban air mobility vehicles in the air tomorrow. In US commercial aviation, our standard of safety is one passenger fatality in seven billion operations. Compare that to about 350,000 deaths on the roads in the 10 years to 2017.

The ground sector could learn some valuable lessons from commercial aviation, even if it doesn’t take to the air. Our safety record speaks for itself, particularly when compared to automobiles.

We did this in large part through safety culture and risk management.

With or without automation, humans inherently make mistakes, and there are different ways to manage the risk.

For the aviation industry, the FAA has a Just Culture approach. Our assumption is that pilots, flight attendants, dispatchers, mechanics, air traffic controllers and others overwhelmingly want to do a good job, so if they make a mistake, we don’t bring down the hammer as long as they tell us about it in a timely manner and it’s not egregious.

For manufacturers, we’ve moved to performance-based safety—we don’t really care how you make a part as long as that part meets an approved standard.

For airlines, we’ve introduced a compliance program that gives us the data we need to make smart decisions, and gives airlines the ability to self-assess and make corrections. Unless they’ve done something egregious, we’re not going to penalize them. Instead, they share the data with us and together we identify the risk and figure out how to solve the problem.

That’s not the land-based enforcement mode right now. You wreck your car, you deal with lawyers.

But it doesn’t have to be that way. If we have autonomous cars full of computers, transmitting data (just like our aircraft), there will be opportunities to smartly figure out how to make the system safer. We know that pilots have gotten to the place where they’re willing to disclose an incident, but will drivers?

If we really want to change the safety of land-based transportation, we have to be able to identify the risks.  I think companies will have to stop competing on safety- but share data to find means to reduce risk. That means drivers, mechanics, manufactures and others who will be willing to tell us about their experiences and having a means to capture all the data that will be available.

As I noted earlier, the aviation sector is not perfect.  But we have a commitment to do safety in a certain way- always open to continuous improvement- and working with the industry to achieve the next level of improvement. The inevitable introduction of new technologies, many from the non-aviation world, will force us to keep rethinking risk mitigation.

That’s why we at the FAA are not sitting around waiting for the next big thing to zip by and leave us in the dust. We actively seek out – and want to be informed about – innovations in aerospace or other industries that we might more quickly and efficiently integrate into the national airspace system.

For example, we’ve reached out to the urban air mobility and High Altitude Long Endurance UAS communities in order to work with them early and often. We want to be on the front end of industry’s evolution, forging new partnerships rather than lagging behind because of our legacy development and acquisition processes.

My boss — the FAA Administrator — has challenged the agency to think bigger, work better, and be open to new technologies, new entrants, and new ideas, especially those that are new to aviation field.  That’s why I’m here, and that’s why the FAA is committed to breaking down the business-as-usual approach to welcome new ideas and possibilities that could radically change the future of transportation.

When we see those innovations coming at us fast as lightning, there will surely be different opinions on whether any particularly technology will help or hurt – just like my sons’ impression of the storm that night.  The FAA needs to find that middle ground that protects the public but allows innovation and opportunity to move forward.

The kind of conversation you’re going to have at this workshop is vital to our efforts.  As you all ideate today, I hope you’ll think about a few questions.

First, what are the kind of skills and knowledge the regulator is going to need?  Are there areas the FAA should be recruiting or focusing on to anticipate these new technologies and entrants?

Second, how might we manage new technologies with very different life-cycles and development paradigms?  I suspect we will not have a decade to develop and certify a new aircraft.

Finally, how do we manage risk and safety in with increasingly open architecture and networks?  We’re already moving down the path of public private partnership in provision of services to drones in the UTM.  I suspect this will only expand- and could be a longer term signal in decades ahead of relying on larger networks for communication, navigation, and surveillance.

I know there are many other questions.  In fact, I’d be the first to admit that in some of the coming users and technologies, I not only don’t have solutions, but I’m sure I’m asking the right questions!  Hopefully, you all can help us with that.

Thanks for inviting me and here’s to an exciting, risk-free day. Now I’d be happy to hear from any of you…