25 thoughts on “Pilotless Air Cargo”

1. I haven’t spent five seconds believing that you are making the argument in favor of this, but…

   Automated airliners? Who the hell is going to offer flights on automated airliners? “Actual safety level” is merely the start of a short, spirited discussion, which will not likely have to be repeated.

   “Pennington says one of air traffic management’s major challenges for the near future will be managing UAVs, because they will be used in a significant proportion of civil and military operations. “

   And then some. This stuff is really cool, but let me help Steve Pennington out a little bit–there’s a difference between civil and military, and as long as the stupid smelly American public gets a vote, there will remain a difference between civil and military. Things which are mundane for the armed forces are unacceptable to civilians, and that includes shutting your mouth when told that the system has been tested, and get on the goddamned plane. Civilians have a tad more autonomy, and too many people demanding refunds on their tickets will run an air carrier out of the skies, permanently, tout de effin’ sweet.

   We’ve already had a mysterious loss of a UAV in Iran. That’s with the level of attention to the mission the Air Force provides, and the Air Force doesn’t have the sort of bottom line airlines do. You’re going to love what happens when a pilotless air freighter loses its’ tether in congested civilian airspace. You’ll hate what happens when a pilotless airliner does.

   With respect, I’m not sure NewSpace is a useful comparison. I’d buy a ticket, given the opportunity, but I don’t think that people willing to buy a ticket, if they could, are a large enough cohort of the population to sweep aside issues of liability. An cybernetic pilot flying the craft, with a warm body in a pilot’s seat, beats a cybernetic pilot with cash customers wondering if this was such a smart idea when a buzzer they’ve never heard before sounds sometime in the ascent phase. Armadillo had better check itself before it wrecks itself.

   We have not yet reached a point in the development of the human race where people have as much need for lawyers as do computers. Whatever one thinks about lawyers, people still deserve to have their interests looked after. Pilotless civilian flight is a remote dream. I can hardly wait to see news footage of passenger drone operators striding through an airport clad in airline captains’ uniforms, the way one can currently see the sad video clips of commissioned Air Force officers sitting in a trailer with a coffee cup on the console, wearing a flight suit. That is a sad thing to see, I had hoped a service with a glorious history, second to none, would have had more stinking pride than that, donning a flight suit to board the trailer. Nose art for the trailers will be next. Kill marks. Survival pistols in shoulder holsters.

   That reminds me–airmen, or spacemen, are cool, they’re heroic. An unacknowledged attraction of aviation, and I include NewSpace in this, is the heroic feeling of it all. Cramming the lowing herds of revenue units into their transportation unit, or smiling at them and saying, “you’re going to love your trip into space!” as you close the hatch on them, will sort of bring that all into focus.

   I started out poo-pooing the poo-poo, and then got somewhat emotional. I could have boiled all this down to, “Nuh-uh, ain’t gonna happen.”

   You’re a good sport to put up with this, thank you for having a comment section.

   1. Just to address one portion of your long post: nobody with a lick of sense, or any experience in the UAV field, puts any credence to the Iranian claim that the RQ-170 was spoofed into landing in Iran.

      1. But land it did, nevertheless. On purpose? No?

         Too bad my lack of experience in the UAV field renders totally invalid my observation that the very first time something like that happens with a pilotless airliner, you know, the kind that has fare-paying passengers aboard, who can be presumed to have entered into a contract to be delivered alive to the destination printed on the ticket, you will see the merriest hell unleashed from all points of the compass on whatever outfit is found to be responsible. No one will book a ticket on DroneAir ever again, and the line building the aircraft will shut down, never to reopen.

         Actually, thinking about, my lack of experience in the UAV field has nothing to do with a dead-certain prediction about the consequences of a pilotless airliner crashing. I’m pretty confident about this one.

         1. No it didn’t land; the photos clearly show that it broke into at least three parts, and it appears very likely that the underside was severely damaged.

2. FAA has done some godawful things in this arena, like shutting down people who fly cameras for profit on RC aircraft. They don’t say a thing if you dont charge for your pictures. How is that addressing safety?

   1. It addresses safety indirectly, and in a manner compatible with the existing regulatory and statutory regime. Flying cameras for profit on RC aircraft means flying RC aircraft over things people will pay for pictures of, which in practice mostly means other people and their stuff – and in the company of manned aircraft attracted to the same things. RC aircraft being flown for recreational or other amateur purposes, mostly fly over unpopulated and uncongested areas.

      It isn’t a perfect correlation, and “no flying RC aircraft over populated areas until adequate safety precautions are in place” would better serve the interests of safety. Presuming, of course, everyone understands and accepts definitions of “populated areas” and “adequate safety precautions”.

      As Rand notes, we’ll probably get that in place in 10-15 years. Until then, the FAA’s existing regulatory regime includes a whole lot of “If you’re not trying to make money off of this, we won’t micromanage your operations on safety grounds”. This is what allows you to fly RC model airplanes without a license.

      Flying RC airplanes for profit, that’s pretty much always been illegal. That won’t change this year. If it is an issue you care about, this year would be a good time to get involved in the effort to craft better regulations ten years down the road. If you want to complain about why it isn’t legal right now, you ought to have been working on that problem ten years ago.

      And that part isn’t unique to RC airplanes. Anything that is presently illegal, that you think you might want to do in the future, you need to get started on making it legal at least ten years before you plan to do it. The FAA in particular is surprisingly reasonable about working with the affected communities to keep the regulations up to date, but no bureaucracy will ever respond well to, “I just discovered that your rules are unreasonable – change them immediately!”

3. Please note that this would be Fedex and UPS, not American.

   1. I know that. I was extrapolating to the inevitable next step.

      1. We could easily start a pilotless airline right now, if we introduced an even bigger innovation: virtual passengers.

         I’m filing a patent tonight…

      2. Sorry, that was supposed to be a reply to Mike James.

4. I’ve got to remember to save my comments before hitting the Post Comment button. Lost another one to an unknown script error.

   Simply put, it’s difficult enough to operate UAVs in relatively empty airspace like over Afghanistan (and Iran). It’s another thing to operate airliner sized UAVs in busy airspace. While the FAA is working (slowly and over budget) to implement the next generation air traffic control system (NextGen), it’ll be a while before they’ll be able to seamlessly work UAVs into the traffic flow out of, in between and into busy airports. As we saw with the loss of the stealth UAV over Iran, command and control of UAVs is far from perfect. There’s also the issue of establishing the proper data links to relay aircraft telemetry from anywhere over the Earth’s surface into the control centers. Freighters range in size from Cessna Caravans (single engined turboprops) all the way through 747-8Fs. There are hundreds (or more) of them in the air at any given time. That’s a lot of data to relay and a lot of vehicles to control.

   1. The data rate requirements for the uplink and downlink are small, assuming true UAVs (the legacy Teledyne Ryan approach, now part of Northrop Grumman) and not RPVs (General Atomics, at least for takeoff and landing). For terminal area operations, you would use line-of-sight links; using modern frequency sharing techniques, you could easily control dozens if not hundreds of UAVs without needing a lot of discrete frequencies. Up and away, all you need is a ping from the UAV once every 10 seconds or so, with a few tens of bytes of status information. If the UAV detected a fault, you would get a burst with the appropriate status info. Not a particularly big deal.

      1. Today, ATC uses voice communications over AM radios to issue directives to aircraft. They’re looking at going to digital data links and that may well be the standard 10-15 years from now (if not a bit sooner). The issues come with integrating legacy aircraft with unmanned aircraft into the traffic flow at a busy airport. It’s a non-trivial exercise for today’s aircraft and it may get more complicated with a mixture of aircraft types.

         I’ve watched several documentaries on US aircraft carrier operations. Despite all the modern technology that exists, when it comes to the movement and provisioning of aircraft on the flight deck, they still use a system of cutouts, washers, nuts and bolts to keep track of everything. Every attempt at developing a computerized system has failed because they couldn’t develop one where the operators could enter information fast enough to keep up with the operational tempo. Going to digital data links to aircraft may have similar issues. It’s one thing to communicate to a small number of aircraft at once. It’s another to keep up with 60+ landings per hour with takeoffs mixed in for good measure. Then there’s the issue of taxi control when the planes are on the ground and C2 becomes challenging.

5. I’m not a pilot, but is flying an airplane really all that much harder than driving a car?

   1. I’m a faily low time (600 hours) instrument rated private pilot. Learning the basics of flying really isn’t that hard. You can go from complete beginner to gettng a Light Sport license in about 25 hours. A private pilot’s license takes a minimum of 35-40 hours but I think the industry average is actually over 60 hours of flying time. Getting an instrument rating (to fly in bad weather) was considerably harder.

      To get your license, you need to pass a written test and a flight test. The written test covers things like basic aerodynamics, navigation, regulations, aircraft systems and the like. The flight test covers flight maneuvers (takeoffs and landings of various types, stall recovery, steep turns, etc.), emergencies, radio procedures, navigation and airmanship.

      Learning to fly means being able to work in the vertical dimensions in addition to the horizontal. That’s why so many people struggle to learn how to land a plane (takeoffs are much easier). It isn’t difficult to learn the basics but you can spend a lifetime trying to master flying. It’s fun. It’s expensive. Fair warning: flying is about as addictive as crack.

   2. It can be.

      Under perfect conditions, it’s easier–no curves in the road, no deer jumping out in front of you. They already have autopilots that can fly an entire route from takeoff to landing without human intervention, as long as nothing goes wrong.

      In bad conditions, or when something breaks… well, there’s a funeral next week for an incredibly-skilled, lifelong pilot and one of the best people I’ve ever known. In a car, you always have the option of pulling over and stopping when conditions get too bad to drive, or the engine sputters and dies, or a sensor gets stuck. In a plane… you don’t.

   3. Is flying an airplane that much harder than driving a car?

      Depends – do you drive like Sandra Bullock? If you are flying a light airplane, you absolutely must not slow below fifty miles an hour, or you will die. Aside from that, there’s some distinct technical detail, but nothing insurmountably difficult. That one thing, though, it makes a difference. It means every single problem you face, you have to solve literally on the fly, at speed. Takes ten minutes or so of work to safely stop an airplane, during which time you have to simultaneously fly, navigate, and troubleshoot.

      Most people can handle this. Some can’t. Computers, basically won’t even try. You can split the effort, have a computer do the routine stuff while a person handles the problems, and you can maybe even put the person in an office on the ground, but if it’s important not to crash you really, really need a man somewhere in the vicinity of the loop.

      1. You really don’t know what you’re talking about here. Modern UAVs (e.g., Global Hawk) do a huge amount of self-monitoring for faults; they’ve been doing the “simultaneously fly, navigate, and troubleshoot” for more than a decade. A Global Hawk does the equivalent of a Cat-3 auto land on every single landing, plus self-monitoring to wave itself off if an unsafe condition exists. There have been a few accidents involving Global Hawks; every single one of them was because of a screwup by either a human operator or maintenance person.

      2. Google has their autodrive car and it seems to be doing fairly well.

         Computers can be programmed with a fairly extensive set of rules to follow to fly an airplane. But yeah, you can’t just turn off the engine and pull over to the side of the road. So, the question is just what sort of problems you might encounter. One thing that helps survivability is flexibility–having two or three different ways of doing the same thing (elevators, trim tabs, engine thrust maybe) and being able to use an alternate method when the primary stops working. So maybe the question is not so much whether the computer can be programmed to fly the plane (we already know it can) but whether the computer can be programmed to fly when things go wrong, or at least to keep the airplane in the air long enough for someone on the ground to figure out what’s going on.

         You mention (kind of) increasing the computer load, and just having the person for emergencies. That tends not to work out too well–airliners might already be overautomated. People are lazy and don’t context switch all that quickly, and if someone doesn’t have to pay attention to fly the plane they probably won’t. There have been trans-Atlantic flight where the entire flight crew has been asleep–if there had been an emergency it seems unlikely that they could have gotten enough situational awareness quickly enough to have done something effective.

         1. Daver,

            [[[There have been trans-Atlantic flight where the entire flight crew has been asleep–if there had been an emergency it seems unlikely that they could have gotten enough situational awareness quickly enough to have done something effective.]]]

            I seem to recall reading somewhere that might have been a factor in the Air France crash in the South Atlantic a while back with only the co-pilot monitoring the plane.

6. Of course, we’ll be doing some market experiments in suborbital in the next few years. Virgin Galactic has redundant pilots, XCOR has a single-string one, and Armadillo is automated.

   And is Masten chopped liver or something? 😉

   ~Jon

   1. Is Masten planning passenger vehicles any time soon?

7. Robot planes may not be much more challenging than robot cars, but the payoff for civilian use isn’t nearly as great. Civilian aviation has an extremely low accident rate, while cars kill tens of thousands and maim even more. Pilots aren’t a major cost driver for airlines, while we lose millions of productive man-hours every day by forcing people to serve as their own unpaid chauffeurs.

   Robot planes would be useful, but robot cars would be world-changing.

   1. Jim,

      Even more world changing would be the convergence of the two into automated flying cars. One of the big barriers to the personal “copters” and flying vehicles featured in so many Sci-Fi stories of the 1940’s/1950’s has been the difference in skill level between pilots and the average driver. Robotic flying cars would be one answer.

      1. I was reading a thread several months back (was it here?) about electric personal VTOLs (kind of like http://www.popsci.com/technology/article/2010-01/nasas-puffin-aircraft-stealthy-one-man-vtol-aircraft ); one of the problems with these is that they’re only good for short hops, which means primarily for use within a city, and dealing with congestion and lousy pilots and not crashing into a building when things go wrong are going to become pretty important. Computer controlled flight might address the first two, having designated crash zones might help with the last (computers might help there as well–the computer could decide that it’s not worth trying to save the life of the passenger and decide it’s better to aim for something cheap and not likely to sue the parent company).

Comments are closed.