17–21 June 2019
Hilton Anatole, Dallas, Texas

Tips from the FAA’s drone pioneer

Posted: 1 May 2017


Jim Williams took over the FAA’s Unmanned Aircraft Systems (UAS) Integration Office one month after Congress directed the agency to write rules to allow drones to fly in U.S. airspace within four years. Although many drone developers were frustrated by the slow pace of the FAA’s regulatory process, people who worked closely with Williams give him credit for helping to draft regulations that have opened U.S. skies to hundreds of thousands of drones that weigh less than 55 pounds, flying within view of their operators and below 400 feet. Contributor Debra Werner talked to him.

Q: A lot has happened in last year or two.
You can say that again.

Q: What has to happen to integrated unmanned aircraft into the national airspace?
I would argue they are already being integrated as we speak. FAA rules that came out last summer started the process. At low altitude, you have commercial and personal drones flying all the time. There have been a lot of reported sightings of drones where maybe they shouldn’t be. That is, I think a little bit overblown, because pilots are notorious for not identifying objects very well. And I think a lot of the sightings will never be validated because there is no way to valid them. Drones don’t show up on radar. But the good news is we have not had any reported collisions or any reported accidents where drones have caused serious injury or collisions with aircraft. That’s the good news. Even with the proliferation, we are still maintaining an excellent safety record. So there is already integration happening at lower altitudes in uncontrolled airspace but more and more aircraft are being approved to fly near airports for commercial operations by the FAA. There’s a whole bunch approved to fly at night. It’s really moving forward.

Q: What has to happen for drones to operate beyond line of sight?
The big impediment there is that the aircraft are going to have to be approved. The initial rules the FAA came up with essentially mitigated any potential problems with the aircraft by restricting when and how they can be flown. The FAA’s assumption is that at any point in time that aircraft could stop working and fall out of the sky, so you need to have operating rules that protect against that.  As soon as you want to fly over people or you want to fly beyond visual line of sight, the FAA’s position is the aircraft has to be approved. There are a whole bunch of ways to get the aircraft approved, but the bottom line at this point is the only ones that have been approved are some military surplus aircraft. Those were approved while I was at the FAA and they are still operating in various functions. They are very restricted to where they can go. In fact, their approval is called restricted category aircraft.

The big barrier that everybody is waiting to see is that first commercial aircraft approved. There’s a company called Aerovironment that builds the Puma and the Raven and the Wasp for the militaries around the world. The FAA publicly announced they have agreed on a certification basis for their aircraft to get design approval. That was a huge step forward. Now it’s up to the folks at Aerovironment to demonstrate they meet those rules they negotiated with the FAA. They are in the process of doing that. Once that happens, things are going to start to open up. It still remains to be seen what sort of limitations are going to be placed on that aircraft once it’s approved. I assume it’s going to be approved for flying over people and flying beyond visual line of sight, at least in remote areas. When that happens, I think you are really going to see applications.

Q: Like what?
There are thousands of miles of linear infrastructure, everything from rail lines to pipelines to powerlines. Regulations require those to be inspected. You could do it with a Puma. Right now, a lot of them use rotorcraft to do that. The railroads actually use trucks with people in them on the rails to do their inspection. They have to shut the rail lines down while they are inspecting them. There seems to be a lot of interest and a lot of companies investigating it but until they get an approved aircraft, they can’t get moving forward. That is the biggest impediment at this point. The initial standards that can be used to approve these aircraft are out there. Things are starting to fall into place. In the next couple of years, I think you will start to see beyond visual line of sight applications in remote areas.

Q: Are there technologies that still need to be developed, like sense and avoid?
Sense and avoid has been developed, demonstrated and approved. The problem is that the initial technology, like so many initial technologies, are fairly crude and very expensive. The one NASA demonstrated last year requires a pretty large aircraft because it uses military grade phased array radars to detect the aircraft in the environment and feed that information back to the pilot so they can avoid them. There are other technologies that are being explored and eventually they will be certified. This is typical of how the aviation industry works. Things are developed in the military and proven out in the military and they migrate their way into the civil market as the cost comes down and the availability comes up. That’s the same situation we have with detect and avoid.

There are a lot of people out there experimenting with different solutions, everything from LIDAR [light detection and ranging] to acoustic sensors and stereoscopic visual. That’s what Intel has demonstrated. They sent one of their quadcopter drones to a forest and told it to fly to the other side. It found its way through the forest without running into anything using stereoscopic vision. Then there’s the radar solution the military and NASA developed. What is happening is the natural evolution of technology is going to take the need and turn it into systems that can be deployed on multiple size aircraft. The technology is proprietary so I don’t know what it is but Aerovironment, as part of their certification, is going to have some form of detect and avoid on their aircraft. That is a key piece. The technology is there, but getting a system that can be approved by the FAA is really the next step.

Q: Is detect and avoid a better term than sense and avoid?
Detect and avoid is the term being used internationally in aviation.

Q: Are there issues with command links?
The real problem is not creating a link. It’s creating a link the FAA would approve for beyond visual line of sight. The current crop of drones use bands that are available for public use. Like WiFi. Both the 2.4 and 5 gigahertz signals that are out there are being used for the commercial off-the-shelf visual line of sight type aircraft. The problem with those links is they are limited in range and the range is not deterministic. The way these systems work — since everyone is transmitting on the same frequency and there is no deconfliction of the signals — they all transit at once and try to rely on the fact that everything is a little bit different in distance apart and so therefore the signals arrive at different times. That works when there are not too many transmitters. But as soon as you get a stadium full of people or an outdoor concert and everyone has their cellphones on with their WiFi transmitters going, all of the sudden that interference goes up. So your range will go down. That makes the range of the signals nondeterministic, which is something the FAA doesn’t like. If you are going to depend on that link to help you avoid collisions, you have to depend on that link to the level that you would be able to assure that you would be able to avoid a collision if one arose as you were flying the aircraft. There are solutions cropping up out there that are not quite commercially available but will be shortly. There is some work going on in satellite and terrestrial systems, but they are not all out there and deployed yet. Again, that will depend upon you having a customer, people who invest in putting the systems out there to allow them to communicate. The technological solutions are resolved, but the implementations of the solutions and getting approval by the FAA are the challenges being worked on now.

Q: It sounds like once things are developed and approved it will take time for them to become small and inexpensive enough for widespread adoption.
Initially the approvals will be fairly conservative. They will be remote areas. They will be of limited duration. But as the experience grows and the systems mature, the FAA will expand the approvals. That’s always the way it’s gone with aviation in general. It used to be you couldn’t fly over the ocean with only two engines. You had to be within 100 miles of an airport with only two engines. Now there are airplanes flying four or five hours from the nearest airport with only two engines. As the technologies are validated and demonstrated to the FAA, they get more and more permissive with the use of them.

Q: What does the FAA’s NextGen air traffic control system offer unmanned aircraft?
For aircraft that would be operating at higher altitudes, it offers a lot of opportunities. The new system that is being installed will have a lot of better automation to deal with unmanned aircraft. For example, with the existing system, you can’t have a flight plan that lasts more than 24 hours. It rejects it as an error. The high altitude, long endurance aircraft that a lot of people are talking about, Facebook and Aerovironment, are demonstrating those technologies so you could have an aircraft that would deliver communications services. Obviously, their intent is to have those up there for weeks if not months. So the FAA system is going to have to be modified. That is one of the improvements coming with NextGen. There are lots of examples how the better, smarter automation in NextGen is going to enable a lot of these unmanned aircraft missions that have been proposed. They are not there yet. The FAA NextGen organization is taking them into consideration and rolling those requirements into future systems. It’s on the radar if you will, pun intended.

Q: Would NextGen make it easier to track manned and unmanned aircraft?
It’s not so much that it’s easier to track them. Customs and Border [Protection] and the U.S. Air Force operate drones in the system with manned aircraft every day in the U.S. and they have come up with ways to work around these problems that crop up in the operations. NextGen just makes it easier, makes it more seamless, better integrated into the system.

Right now, Customs and Border flies up and down the border every day at 18,000 to 20,000 feet altitude under the positive control of FAA air traffic controllers. As we speak, there are probably two or three of them up there flying the borders north and south. It’s doable, but the FAA limits the numbers, how many can be up at once, how many can be in one area at a time, because if they do lose the command and control link, things get really complicated. For example, in the current communication system those drones use to talk to controllers, there is a VHF radio installed on the aircraft flying in the system. But the way the pilot talks to the controllers is their voice has to go through a satellite in geostationary orbit down to the drone and out the radio to the controller. Then the controller’s voice goes back the same path. If the pilots lose their connectivity to the drone, they have suddenly lost their ability to talk to air traffic control. The NextGen systems will include a way for that pilot on the ground to connect directly to the controller through a ground-to-ground system instead of having to go through the aircraft. That way if they did lose connectivity, the pilot and the controller can still talk and the pilot can explain to the controller what the aircraft is going to do and how likely they are to recover communications. It is things like that that are going to make it easier to operate in a more integrated approach going forward. That is part of the new communications system being installed as part of NextGen.

Q: Isn’t NextGen also focused on data transfer?
Yes. Data communications will also be implemented as part of NextGen. Right now, the way that system works where it is implemented is data goes over a ground network to the radio that transmits it to the aircraft and the pilot onboard sees it. If you are going to do the same thing with an unmanned aircraft, again, it would have to go back through that satellite communications down to the pilot on the ground. In the NextGen environment they would simply route that data communication directly to the pilot on the ground and never go through the aircraft. It’s a lot more reliable and a lot faster. All of that is included as part of the build out of the NextGen systems as they go forward.

Q: Do you picture many more unmanned aircraft sharing airspace with manned aircraft?
I think so. There is definitely a business case to be made if they can overcome the technological challenges for long endurance aircraft providing communications services. There is never enough bandwidth. There are always going to be places where it is not economical to put up a cell tower. There are places in the world and even in the U.S. where it is difficult to get Internet at all because there is just no connectivity. So having the ability to put up what amounts to a cellphone tower in the sky to enable communications both data and voice, I think there is money to be made there. The real challenge is getting aircraft that can fly for long periods and still have enough power onboard for an airborne transmitter. People have made solar airplanes that can fly indefinitely. The problem is they are very light and fragile and they can’t carry much. The challenge is getting one big enough and strong enough to carry a significant payload for months on end.

Q: That is one application. But shouldn’t I picture thousands and thousands of unmanned aircraft sharing the U.S. airspace in a decade?
Not at altitude. I think you will see thousands and thousands at low altitudes, flying over power lines, pipelines, rail lines, delivering packages, etc. The thousands and thousands will all be in that low altitude airspace because that’s where the economics are for spending the money to buy the aircraft to set up the services. Other than what the military does, there is not a whole lot of demand for aircraft operating in that same airspace where the airliners are.

Q: What about FedEx and UPS?
That is theoretically possible. But to them it’s all about the bottom line. Given that you are still going to have to have pilots on the ground and you will have the extra technological and cost challenges of maintaining a link between the pilots on the ground and aircraft, at this point it is not economical to replace a manned aircraft because pilots these days don’t cost them a lot of money for those kinds of flights. In the long term when computers can control the aircraft and you can have multiple aircraft monitored by a person who may not even be a pilot, that is theoretically possible, but that’s a long way off. The technology is far from being able to fly without a pilot in the loop and the regulatory structure just doesn’t support it right now. One of the issues that is hanging out there, even over the driverless cars, is if you are riding in your driverless car and it accidently runs a stop sign and hits somebody, who is liable? You or the guy who wrote the software that screwed up and allowed it to run the stop sign? There are some really complicated questions there. Right now, the pilot is ultimately responsible for the aircraft, from an FAA standpoint, regardless of where the problem was. The Amazons and Googles of the world are working to resolve that by trying to create a means of controlling multiple aircraft with just a human overseer. When that comes along, which is years away, then I think you’ll see FedEx and UPS start to talk. Because if they could have one person looking over 10 aircraft, now all of a sudden the economics become more realistic.

Q: When there are thousands in this lower altitude, below 400 feet, will they be part of NextGen?
They are operating in Class G airspace, which goes up to either 700 feet or 1,200 feet depending on if you are close to an airport or not. The FAA doesn’t exercise any control whatsoever in that airspace. As far as the FAA is concerned, you are on your own. That is the airspace in which I think you will see a lot of operations in the near future. Right now, it is up to the pilot to see and avoid the manned aircraft. As the technology matures and the aircraft can do it itself, then you’ll see people flying beyond visual line of sight in that airspace.

Q: When you were at the FAA you established an approval process for drones.
Those are expiring and starting to drop off and everybody is starting to use Part 107. The number goes up every day but over 10,000 pilots have been approved under the new system. All these numbers are available from the FAA. They keep track on their website. One number I thought was most interesting was that there are over 600,000 individuals like myself registered to fly their recreational drones. That is more than double the number of actual licensed pilots who fly manned aircraft in the entire country. That is just in the first year that the registration was out there. I think it shows you that everybody now has access to fly in the atmosphere, whereas in the past, it was only the very wealthy or the very determined who could do it. And it’s really expensive to own and operate even a very small, low-end general aviation aircraft.

Q: What do you fly?
Recreational types that are more toys than they are expensive ones. I’ve resisted the temptation to have the DJI Inspire because they are really cool. But I don’t fly them enough to make it worth spending the money to operate one. The people who use those are really avid photographers or doing it for a business. Basically, what they are is a flying camera platform. I fly for the fun of flying the aircraft. The toys are more fun to fly because you have to work hard to keep them from running into stuff.

Q: It sounds like I shouldn’t wait for UPS to deliver my packages by drone?
There are real regulatory challenges to doing that. UPS is an interesting case because they are all about increasing the number of deliveries each individual driver can make so they are all about saving time. They are only interested in delivering packages via drone if they can save a significant amount of time. That will allow them to amortize the cost of the ac and the installation. It gets complicated because your driver becomes a remote aircraft pilot. Does that mean you are going to have to pay him more? If you are going to have to pay him more, you are giving up that benefit of time you would save. For the economics of it, you have to peel back the onion. It looks cool for the brown truck to stop at the end of a long gravel road to somebody’s house and instead of driving up that driveway, zip the drone over and drop off a package and zip it back. But there’s the cost of the aircraft, the cost of maintaining the aircraft, the cost of training the pilot, and the driver has to sit there and watch it the whole time. What UPS would like to do and they demonstrated it in the press recently, they want to send the aircraft to deliver the package to person A and have it meet the truck when it’s delivering at person B’s house. The driver stops briefly, sends off the drone and keeps driving. Now, that pilot’s time is being used to drive the truck instead of flying the aircraft. The problem with that is it’s a much more difficult approval to get from the FAA. The aircraft has to fly on its own. That’s going to drive up the cost of the aircraft. It’s going to have to have sensors that can detect a helicopter flying by and avoid it. And all that would have to be approved by the FAA. Whereas the first scenario, where the UPS driver just sits or stands by the truck and flies the drone out and flies it back, they can do that today. That’s allowed under the new FAA rules. But the economics of that don’t really add up for a UPS or a FedEx. It’s interesting when you start to look beyond the technology at the business case behind these. That’s why you see a lot of big companies spend a little bit of money investigating how it all works, but you don’t see companies pulling the trigger and spending millions to buy drones for everybody. I think that’s going to happen.

Things like inspections of roofs after hailstorms. For the insurance companies, that’s a big cost for them. I talked to a State Farm person who said when they get a claim of roof damage, they send an adjuster out to talk to the person. They then put in a work order for a contractor to come out, climb up on the roof, look at it and take pictures and determine if the whole roof has to be replaced, if it can be repaired, etc. Then that information finds its way back to the insurance adjuster who can then authorize the repair. You can see a lot of time is going to elapse. Whereas if the adjuster had DJI Inspire in the trunk and was trained to use it and trained to see what a good roof and a damaged roof looks like, he could pull up to my house, talk to me, put the drone in the air, fly over my house and in five minutes he could map it all out, get the results and say before he left whether my roof needed to be replaced or repaired. That is the kind of thing that has potential for huge deployments across the country for one of these big companies, but they haven’t figured out all the logistics of which aircraft to buy and how to train their adjusters. The adjusters would need to get FAA drone operator certificates. The company would have to set up a maintenance and tracking system of all the drones and all the batteries. The logistics of making that happen even though it sounds pretty straightforward really get complicated for a big company like State Farm.


Return to Aerospace America Home 

Return AIAA Home 

Return to 2017 AIAA AVIATION Forum Home 


FAQsContact UsPoliciesMedia