You're not aware? Indeed. LionAir.

What no one has thought to mention is that the present generations of hardware on aircraft are in a lot of cases 20 or 30 years out of date. For a long time, the manufacturers would not use a CPU chip more powerful than the 386, as that was the most recent chip that had been security certified for aviation use.

The technology has moved on massively since the 386, as has been demonstrated by producing hardware that was capable of guiding a radio controlled model across the Atlantic a while back, It was launched and flown to "cruise" height using normal Model control radio, then put into automatic, flew to Ireland, and then was landed by another operator on the ground.

The chips, and the size of the electronics are such that you could have 20 systems capable of performing that task in a box the size of the normal FMS on the aircraft.

The same restrictions of performance apply to many other computer systems used by the automation of the aircraft, so as a result, there are significant limits to what those systems can do in terms of analysis and response, not because the programmers can't improve the programs, but because the hardware is not fast enough or large enough capacity wise to handle all that could (and should) be programmed into the systems.

I've worked with computers for over 40 years, and also have a ME CPL/IR, and have worked on detailed and in depth research on ways to develop and improve the performance and capabilities of in flight systems for commercial aircraft, and I know that with the right funding, to make sure that the hardware has been properly quality assured, and with the right people working on the software side, the improvements that could be made to aircraft systems go a long way towards making the operation of even complex aircraft very much more straightforward than things are now.

The most damning phrase I have heard from pilots is "what's it doing now?" That phrase is a reflection of 2 issues, the lack of a good human interface between the computers and the crew, and the lack of in depth understanding by the crew of how the aircraft they are flying really operates, and how its systems interact.

A lot of the lack of crew knowledge comes down to the reluctance of bean counters to let them learn, because they can't see good reasons for having that level of knowledge in order to deal with 0.001% or less of incidents. That might be because the bean counters are never anywhere to be seen when an accident occurs, they are very good at distancing themselves from difficult situations.


Right now, I would not be happy with a single crew commercial aircraft, simply because the human interface with too many of the disparate systems is so poor, and tortuous, and things like ATC are still too much of a distraction from the fundamental task of aviating.

If the interface systems were improved, and the hardware performance increased appropriately, then I might be more inclined to consider single crewed commercial flights, I flew enough single crew IFR to know that it's not impossible, and also to know that the automation is not yet capable of providing adequate support to a single pilot to allow safe operation in all flight conditions. There were times when just flying the aircraft single crew was enough work load without having to deal with ATC that sort of spoke English with a heavy foreign accent, or ATIS that was computer generated but so bad quality that you have to listen to it several times to get the information needed, or instruments that have a poor interface so take significant time to set correctly, or the issues of trying to set the next comm frequency accurately in turbulence.

I could go on, but I think a lot of people will already have got the point I am trying to make.

If Boeing can produce drones that can be flown from 9000 miles away, in a military environment, it is not going to be long before that same technology can be enhanced to have the redundancy, safety and security needed to allow commercial operations It will probably be introduced into freight operations initially, for testing and proving the capabilities, but it will come for passenger ops, and nothing anyone says here or elsewhere is likely to change that.

I saw and worked with prototype systems that were capable of going way beyond the way that even current systems are working, and it was exciting to do it, and if I had the money, I'd still be doing it, and possibly testing some of those developments in light aircraft, just to prove that they can and do work.

Piloting has changed massively in the last 30 years, I've seen it happening, and it will change massively again over the next 20, of that there is no doubt.

Ideally, the good pilots that are involved now would be closely involved in making sure that the next generation of avionics and systems are a significant improvement over what has been done so far, as too much of the present systems have been designed by non pilots, with pilots then expected to change how they work to suit the systems. That's not good, the automation has to complement the pilot, not oppose the pilot.

When that happens, then everyone will be happy.

Amuses me that there is talk here about increased automation and the concept of only one pilot,yet we can't seem to devise a cast iron system which will ensure that an a/c does not depart with the engine cowls unlatched.

Regarding passenger perception of automated flying, if the choice was a flight costing $250 vs one for $200 without a pilot, you'll find the majority will choose the one costing $200 (you'll probably find surveys that would sway most people with $5 difference). The majority masses do not care about the types of aircraft they fly, the boarding process, the crew competency, they just care about the price showing on their computer screen at the time of booking.

Once you have swayed the public with driverless cars I don't think the public will have any issue with 'driverless' aircraft...

But you must have a pilot, otherwise there's no one to blame for not spotting the open latches on the walk round.

The B-2 is quite automated, and flies with a crew of 2. Missions can go as long as 39 hours (limited by engine oil consumption). In low-stress parts of the flight, the PNF will lie on a mat on the floor and nap. Perhaps that scenario would be fine for long-distance air carrier ops, but I can't see single-pilot for these types of flights. Even if the pilot could nap while the autopilot flies, if an emergency arises, the pilot will not be immediately alert. I would surely think it would take one longer to assess and analyze the problem if just jolted awake, versus paying attention all along.

The idea of a ground-based pilot flying (as some military aircraft are now operated) still requires a pilot, getting paid. If the idea was to save labor costs, I don't see the point.

I would not be a passenger on a pilotless aircraft. George might do a superb job flying 99% of the time, but computers do not have the flexibility of the human brain. I can't imagine a computer having the same ability to react to an emergency that a human can. A great example: the Airbus that landed in the Hudson River. With a dual flameout, what would the computer do? Lower the nose and maintain airspeed, of course. Then what? If no runway is within glide-range, can it be smart enough to choose the river, rather than the densely-populated city? Those lives were spared because of an experienced crew that could think flexibly, not constrained to 2 or 3 choices of the computer programmer. (I understood that the crew would have run the engines to max power to save the flight, but the computer prevented that. "I'm sorry, Dave. I'm afraid I can't do that.")

When the inevitable crash comes, with no pilot to blame, who becomes the brunt of the lawsuit? The technicians who allowed something to fail? The manufacturer of a flawed computer? The programmer of that computer? Surely the dead or injured SLF's families won't be saying "well, they did get their ticket for a good discount because pilots weren't having to be paid, so it would be unfair for us to sue".

another great example - Lion Air crashing in Bali with two awake pilots in the loop

there are as many cases of pilots being the problem as there are the kit being the problem

rotorfan : excellent remarks. but what is currently done in the military ( e.g UAVs) cannot be duplicated, and by far, for passenger transport .

Heathrow Harry : Bali : good example among many .
but to automate approaches you need some ground infrastructure . in Bali on that runway there is no ILS, so what do we do to go there in a more automated way ?
landing systems equivalent to CAT3B at the end of every runway at every airport ?
What will be the cost of that ? I think in countries like Indonesia with over 200 airports , it is definitively cheaper to leave 2 guys up front ,( especially if the second one is paying to be there:rolleyes: )

OK ... so now were safer with no pilots. Has someone got data on the times when pilots did "save the day" or "get it right"...."no".... thought so.... :rolleyes:




(but probably about every other flight that leaves the ground)

or get it wrong and its never reported.........

right now the balance is in favour of 2 pilots but, just because we don't like the idea that we can be replaced by machines, is no reason to say it won't happen

When it goes wrong, its usually reported by CNN "raw footage" of a smoking hole. That's about all that's needed.... isn't it..! ... :ok:

Good post Irish Steve, you may well be right, one day, but before that day all security issues have to be resolved with no prospect of a well armed and trained suicidal team of zealots taking over a ground control station or simply setting up a powerful jamming station that can issue rogue commands or, more simply, no commands. Either of these scenarios would be able to cause carnage of biblical proportions so I think the technology may be available at least a century before the security issue is resolved.

All of the above makes interesting reading.
Events like Bali(lion air), and AF 447, make a strong argument for UAV's, because in both cases the crew WERE the problem. AF447 would have recovered from a deep high altitude stall if the crew had of just let go of the controls! And the Bali splash down was just unforgivable.
The root cause IMHO goes much deeper than poorly trained crew. Airline management, bean counters are saving a fortune on training, P2F, actually makes money, but produces an inferior product. Basic high altitude hand flying skills have deteriorated over the years because the automatics do such a superb job. If anyone on the flight deck of AF 447 had of recognised a deep stall........ Well the result would have been completely different.
The regular 6 monthly sim sessions concentrate on V1 cuts, OEI approaches and and go arounds. Basic flying skills are slowly being eroded by the advent of better and better automatics.
The argument against UAV's is made perfectly by the glide approach in the Hudson River. That needed a competent pilot hand flying and capably assisted by a competent First Officer. Sully did a good job, but no better that the vast majority of competent crew would have been able to do. Sullenberger, or anyone of the vast majority of his co-hort would have performed very simarlarly, ie without a human pilot that aircraft and it's complement would have been doomed.

Have you noticed that most aircraft prangs are not related to system failures, that is to say the defect by itself was not solely responsible for the disaster, it required the crew to really stuff things up. The solution to this is simply better trained crew. Some significant time actually hand flying the aircraft in all configurations, and much more emphasis on CRM. When an F/O says "go around" the captain better bloody well go around. Air India Express apparently went for two years without any documented CRM training, and then lost an aircraft to very very poor CRM. Who would have guessed. Garuda lost a 737 at Jogjakarta, under almost identical circumstances, poor CRM.

If we as a pilot body keep repeating this sort of accident we only strengthen the argument for pilotless RPT.

How will ATC work? Will the instructions come from them in the form of a data package that uploads to the FMS via satellite uplink? If that is the case will the FMS pick up ATC errors like pilots do day in day out? ( just as ATC pick up pilot errors). I find that hard to imagine, ie when ATC issue an instruction to aircraft A instead of aircraft B it is usually picked up by the pilot and they say something like " confirm left heading 030 for Delta 116?" how will the FMS deal with that? Will it just accept the instruction?

Further to my last post, the point I think I was making is that we won't have removed the human element from the system, it will still be there , just on the ground. Those humans will make mistakes as humans do and computers will do what they've always done and stick to the rule of " garbage in- garbage out" .

TBH I'd far rather they spent the time and money coming up with a foolproof way of telling us exactly what the ACTUAL weight of the aircraft was before take-off

I'd far rather they gave us one rostered day every second month to sit in the classroom and go over all the changes to airways/ SOP's/ ATC /approach requirements/ RNP- AR/ EFB procedures/ Engineering requirements etc etc.

I flew into Marrakech about 5 weeks ago. ATC down there is the poorest I have witnessed by a country mile. As we descended through Fl150 I could see those great big Atlas mountains to the south and all around them heading our way were some of the biggest thunderstorms I have seen in my career, the weather radar was lit up like a Christmas tree.

Needless to say a lot of weather avoidance enroute to Agadir and once again ATC was virtually non existent( no radar in Marrakech).

The workload was high. Very high. The workload management was good. End result: Perfectly safe landing in Agadir.

The thought of me sitting there in a 65 tonne Jet with 200 souls on board doing all this on my own on a regular basis, quite frankly makes my blood run cold.

There's a reason single pilot airliners with hundreds of people on board don't exist.

When will they realise.:ugh:

Actually, you must have a "good" pilot to spot the unlocked engine cowls and not one who's just taking a quick stroll around the outside to "show-off" his gold bars! Bring back the FE... the A319 "incident" would never have happened!

go around flaps15,
Hear, hear!
Regrettably, few, other than professional pilots, understand how workload can build up, and sometimes rapidly.