So all we need is well specified, well designed, well written software, free of bugs running on bug free chips that have a totally reliable power source driven by people or systems that never make mistakes. What's so hard about that?

Meanwhile, every year humans scatter dozens of perfectly serviceable aeroplanes and their passengers across the landscape. But that's ok, because they know what they are doing... :E

Call me a cynic, but when I fly as a passenger I want whoever is controlling the aircraft to have as much of a vested interest in getting it down in one piece as I have. :O

Actually, computer systems can also "not know what they are doing". They are usually designed around the assumption that they have a complete picture of what is going on, and complete control over the process. When either assumption (or both) fails, the computer will depart its zone of expertise rather abruptly. A well-designed car or train control software can often recognize the problem and shut down until help arrives; an airplane might not have that luxury. ("Is there a software engineer on board?")

Is there really a solid economic driver to do away with pilots? Yes, pilots are expensive, but put up against the overall cost of operating a large passenger jet, the pilot's costs are, if not insignificnt, minimal.

Better to look for fuel economy improvements or lower maintenance costs.

Plus I think the psycological hurdle of getting aboard a pilotless aircraft will be a step too far for most passengers.

Mechta - I don’t agree. In fact I’d go along the line that millions of flights have ended incident free purely because there were pilots on board, not in spite of them. Until we grasp what pilots actually do to make up for the deficiencies in aircraft, their systems in the information supplied we haven’t a hope of removing them from aircraft.

The same was being said 40 years ago about flight engineers.

Look what happened to them.

Pilot workload was reduced through automation so that the pilots could take on the flight engineer's tasks? and for those tasks the pilots can't do, they phone home to base?

Passenger jets cost most when they're on the ground not doing anything, rather than in the air, earning revenue. Pilots don't just cost wages - there's the accommodation down route plus the caps on hours. Have 'em monitoring the systems remotely/centrally and you do away with all that, all for a small anti-social hours supplement to pay to compensate for a 24/7 shift pattern. You could also break the idea that the same pilot (or crew) have to fly the entire trip: 'pilot' A monitors/flys remotely the takeoff, then when his shift finishes, hands over to 'pilot' B who does the landing. Pilot A meanwhile has driven the 20 mins home, and is spending time with his family until his or her next shift starts again in 14 hours time (or whenever).

Yes, and the automation was driven, as it almost always is, by economics.

The suggestion that a pilot on board is equally unnecessary is just the logical extension of the same argument.

@DaveReid: no, I believe it isn't. The FE could be made redundant because other humans (pilots) were able to take over their task that automation couldn't take over. If you extend the argument (and the automation), you could maybe also eliminate the FO, but that strategy also relies on pushing some non-automated tasks onto the pilot. In any case, the pilot retains the task of supervising the automatic devices. To eliminate the pilot altogether, you have to a) automate all tasks, not just the "easy" ones, and b) do that to a degree that eliminates the need for on-board supervision, which is presently the fall-back. You can't do it in the same way (reduce workload, reduce the humans) because you have to reduce the workload to zero, moving the computer from its support role(s) into a position of responsibility. That looks like a qualitative difference with unique challenges, and while there may be some economic pressure towards achieving them, it is hard to predict when that's going to happen, if it ever does.

But yeah, maybe in the future we'll see camera drones doing the walk-around check on the aircraft.

Look at the legal arguments around "If a self driving car crashes and kills someone - who is to blame?"

I didn't say eliminate the pilot, I said eliminate the on-board pilot.

It’s not a question of if, it’s a question of when.

Liability issues has been settled in the U.K. for self-driving cars. It will be the manufacturer who will be responsible. This likely will be followed in other jurisdictions.

If you can distribute the cost for two pilots over 400 sold tickets there is very little economic drive... The costs for two pilots in a 19 seater with 12 sold tickets are more significant.
So economically pilotless small planes are more interesting. For cargo planes the psychological hurdle is not relevant and here goes a similar argument but then based on pilot wages per kilo cargo. And getting rid of the pilot allows another 100kg or so extra cargo in the plane.

So looking at current operations, I expect that the small mail runs (<1000kg max load cargo planes) are the first candidates for drone-replacement, but there also is an option to go to smaller (50-100kg load) unmanned planes (that can't even fit a pilot).

All good in theory MathFox, but in my experience small aircraft and freight operators are the least likely to have the inclination to invest heavily in brand new, state of the art, pilotless aircraft. If anything they tend to have relatively old machines and would be more likely to follow the pilotless trend some 20 years after the fact once the pilotless aircraft become available cheap on the used market.

In "2001, A Space Odyssey", there were 2 pilots AND flight attendants. Pan Am, wasn't it.
.

It takes how long to get a brand new design from the drawing board into service? When Airbus or Boeing announce that they are developing an autonomous airliner that would take at least 20-30 years to enter service then I am not going to lose sleep over it. Most of the conversation is based on the false assumption that the two pilots are reading the paper and pressing the "land now" button.

False assumption or merely unkind stereotype? Anyway, the question is whether the paper reading and "land now" button pressing needs to be done at the front of the aircraft, or whether it can, and indeed will, be done remotely.

Seems to me, some of the arguments put forward against the idea echo those of the hand loom weavers, and things didn't come out too well for them.

Mechanical looms had limitations (thread count, colors) over hand weaving. Turns out, these limitations didn't matter so much. We're still discovering what the limitations of computer systems are--in aviations, some of these discoveries have already cost lives.