My view regardless of how close we are to some technology that may be in the is that even if the tech exists, it’ll be along time in my opinion before airlines and passengers are comfortable with having no pilots on board.
Perhaps I am naive, but I really do not see this happening.
Hasn’t the military been doing autonomous flights for almost 20 years now? Looking at the RQ-4 Global Hawk program and the X-47 program, they seem to have it down pretty good. None of this automation really seems like new or ground breaking technology.
And I’m sure there are other military autonomous programs out there that we have never heard of…
Clint,
Those flights are mostly still flown by pilots, just remotely. They also have a rather high loss rate, not all of it due to combat.
Chris
I’m no expert when it comes to this, but through some research it looks like Global Hawks have very limited pilot interaction. The flight plan is uploaded to the aircraft and then a pilot initiates a taxi/take off sequence using a keyboard and mouse from a ground control station. The pilot interacts with ATC but the plane does everything on its own, it even determines when it should abort a take off/landing. And if a pilot needs to change course in flight it’s all done by entering new coordinates with a keyboard/mouse interface. The pilot can enter rates of roll or climb/descent but the software won’t allow a pilot to enter a parameter that will cause damage to the aircraft or put it into an unrecoverable condition.
The loss rates don’t seem to be to bad, with 4 aircraft lost over 300,000 flight hours, and one of those was due to being shot down. Another loss was due to pilots misunderstanding of a nav fault disabling the wrong Nav computers, but they were following the manufacturers checklist.
Let’s call that two crashed then. Two per 300,000 flight hours. United Airlines flew roughly 3,000,000 flight hours last year, so using that same rate would equal to twenty crashes per year. That doesn’t sound “too bad” to you?
It’s estimated there are about 90,000 flights every day. Even if we discount the 2 you mention that would mean 2 crashes every 3-4 days. Not sure that would work for the flying public.
Adam
After digging into the accidents a little more, another was caused by the loss of engine oil while flying over the Atlantic. I’d imagine that you’d have to be a very brave pilot to fly a single engine aircraft over the Atlantic without having any viable emergency landing points in case of engine failure, as happened with the one that crashed off the coast of Spain. So if that was a manned aircraft, I’m pretty sure the outcome would have been the same. Is that chalked up to military risks associated with these flights or bad aircraft design and automation?
The other flight crashed in 2012 near Naval Air Station Patuxent River was due to “a motor failure in the right inboard ruddervator, the movable hinged section of the V-shaped tail surfaces on the rear of the drone” and that the pilot “did not follow established emergency procedures”, however, the report also states that the pilots actions could not be conclusively contributed to the accident.
I cannot find a report or article that directly links these accidents to automation. If anything, human error led to some of these. There were some crashes in the late 90s/early 2000s that were directly linked to automation issues but those were on test platforms, so I’m not sure those should factor into the overall crash rate of an aircraft.
Clint,
I am not sure where all of this information is available. I can tell you that I fly with many former drone pilots and none of them think that the technology is there for automated flight.
Chris
The info I find comes from following sources from Wikipedia and then following more sources from the pages those get referenced to. Eventually I can piece info together or get to the source documents.
I hear different comments from the pilots I work with. I talked about this issue with an experienced RPA pilot who transferred from manned aircraft, flying the C-17, to MQ-9s. His take was that the automation is here and the reliability is here but manned aircraft pilots are very resistant to the idea. There’s a lot of animosity towards the RPA pilots and a lot of resistance to the idea of autonomous flight. He’s even noticed that manned aircraft military pilots don’t consider the RPA guys as being worthy of wearing the Wings that they’ve earned and that they didn’t have to work as hard to earn the same pay. Maybe there’s some emotional responses that we hear from manned military pilots? I’m not sure what you hear, but I think the technology is here and it might move forward faster than we expect.
I’m not in on the whole business model of how airlines work but are pilot pay and benefits really that big of a chunk of the operating costs of an airline? How much would my ticket cost be reduced if you weren’t in the cockpit? I have a feeling, if automation took over flying, I wouldn’t see any reduction in ticket price and instead that money going to an already very wealthy individual. It also seems like the airspace around airports is already congested as it is, would adding more planes into an already busy environment with automation really be the best idea? And could someone hack into these systems and take control of large amounts of aircraft at once? Those are the concerns I have more than the technology being available.
Interesting that you bring up the security issue. One of my current students owns a large cyber security tech company, and he says the biggest hurdle to automated airline flights is exactly that. His company’s current big project is putting together a package to show the weaknesses of current airlines to cyber attacks and how that could easily impact automated flight. Hack the system of an airline that has gone fully automated, and you’ve got thousands of potential 9/11s flying around at once.
No Bueno.
Several points to address here. Very few of the people talking about automated flying are actual airline pilots. They are drone pilots, security experts and very wishful aircraft manufactures. Actual pilots know that flying takes more than just automation. There is a science and art to it. Knowing just how close one can come to a thunderstorm because you can read the clouds, how to handle a cross wind landing, how to handle various emergencies, the people factor, etc are all things that only an airline pilot can do. Judgement is a huge part of flying, that is very difficult to automate or have effectively done by somebody half way across the country.
Clint, I can guarantee you that the military is not publishing all of their crashes, especially those that happen over seas. RPA pilots, at least those that have not flown actual airplanes, are not pilots. You said yourself that the automation does much of the flying, if this is the case then what is the pilot doing? It certainly isn’t flying. Flying an airplane for thousands of miles away is not the same as sitting in it yourself, having your life and others lives dependent on you. Entering commands into a drone is not the same as flying an airplane through a storm, it just isn’t.
I am not sure what percentage of costs pilots make up at an airline, but with captains easily earning over $300,000, it is significant.
Peter, your friend is a tech expert, not a pilot. I would say that he probably has very little idea of what it takes to actually fly an airplane.
When the Port Authority can get the AirTrain at EWR to actually operate correctly, I will get somewhat concerned. I have been based in EWR for fifteen years now and that thing is still a hunk of automated junk.
I would further point out that Boeing has not exactly been doing so well with their new, automated technology recently.
To clarify, his stance is that airlines will not go the fully automated route any time soon, if ever, due to the security risks involved. I’ve always been doubtful purely due to the inherent flight safety risks involved, but his was a new angle I hadn’t considered. It certainly seems the odds are stacked against a wholesale overhaul movement towards automation. It just isn’t feasible in so many ways.