Just yesterday, we were on final approach to Kennedy, and we had to go around last minute due to a large flock of birds that had decided to come by and say hi. That same day, we were over Britain, and it started to get a little bumpy, and I preemptively decided to turn on the seatbelt signs and requested a decent, which was approved. Then, chaos erupted above us, and luckily we only got it for a little bit. All in one flight. And these aren't uncommon, this is stuff, along with much more, that pilots have to deal with. Good luck getting a computer to do all that, having to deal with these idiots flying drones near airports, birds, making the decision to go around, extreme turbulence, collision avoidance, and much more, while dealing with everything else in addition to emergencies. I've been flying for 29 years, and for 29 years I've been hearing about computers replacing humans in the near future, and I'll keep hearing it for another 29 years.

This thread proves the old adage:

What's the difference between a pilot and God? God doesn't think he's a pilot.

Flying is not materially more challenging than driving, from an AI programming standpoint. Airplanes get 1000 feet of space around other airplanes. Autopilot has existed for 50 years, and auto-land is there, though still not as "passenger friendly".

Airline labor costs are the highest non-fuel cost an airline incurs, and pilots make up a big chunk of that. You can believe when the opportunity exists to automate out one or both pilots, the airlines will seize it.

Like others have said, it'll be a drawdown, with 1 pilot and a centralized emergency pilot who is monitoring multiple flights. Then eventually it'll just be one pilot while the plane entirely flies itself. Then, ultimately, that pilot goes to the ground.

It's not 50 years away. At all. It's 20, at most.

barney captain wrote:

Most of you are missing the number one reason this can't/shouldn't happen - judgement. You can't teach a computer "spidey sense" - and it's a player far more than most think.

Just tonight we were approaching the Rockies and felt/noticed just a tickle of what felt like mountain wave. With nothing forecast we queried the controller - nope, smooth rides ahead. As a precaution we decided to ask for a lower altitude and sure enough, a few minutes later at our previous altitude all hell was breaking loose. Our new altitude remained smooth.

Btw, in that same situation in the AI cockpit, at what point in the moderate turbulence would the accelerometers determine (after the fact) that the cabin crew should be seated?

One of an unlimited number of examples of why this is a very bad idea.

Flying is still an art.

If we can (theoretically) replace pilots, why would we need cabin crew? IFE screens can give passengers instructions, and robots can hand out peanuts and pour drinks. But you could count me out as a passenger.

Not just yet, but it is coming and it will be in our lifetime. Its just a matter of time.
Im sure the airlines salivate at the prospect of replacing 2 or 3 pilots with a similar number of "technicians" who will monitor gauges and the flight at a minimal cost.

Gives a whole new meaning to "scope relief."

It’s not coming, you can’t program a computer for every possible scenario as there are limitless issues that can happen. And a computer can’t react as fast as a person in possible catostraphic scenarios.

How will a computer react to a highjacking?

Boof02671 wrote:

It’s not coming, you can’t program a computer for every possible scenario as there are limitless issues that can happen. And a computer can’t react as fast as a person in possible catostraphic scenarios.

How will a computer react to a highjacking?

How would the hijacker take control?

Wait...now you’d have to worry about hi-HACKers. Think about those possibilities.

We have just seen the issues where automation overwhelms trained pilots. Imagine how many times having pilots in the cockpit has saved lives. Do we really want the extra risk to lives in order to save a few pounds on a ticket?

BlatantEcho wrote:

.
It makes me chuckle to see people today argue with this sort of progress.

At what point does progress reach its pinnacle? As we progress, eventually all of our jobs will be automated and we can just sit around and form even more opinions about things unrelated to us, and share them on even faster networks with better features and ....

Call me a romantic, old,or better yet, be modern and google (not capitalized because in the internet age we don't do that anymore) me up a buzzword adjective, but I think we've kinda passed the point of balance between tech and nature.

How will a computer handle an electrical failure? A bomb?

The logistics are insanely complicated and frankly not even safe to consider. A single pilot onboard to provide minimal inputs and react as needed, I could buy that. But fully unmanned, not a chance. You think aircraft will taxi themselves? Provide take off and landing clearance themselves? Take alternative routes due to weather themselves? What about an engine or tech issue? The plane will divert to the closest appropriate airport itself?

Those that comment that it will happen in 20-30 years are humorous. It will take 20 years just to build the infrastructure to try to support something like this. Imagine all the testing and approvals that will need to occur for the entire world to adopt this idea. The technology doesn't even exist yet. You could make a list a mile long as to why this won't happen. I hope I am alive to find this thread in 20 years and see where we are then.

But who then will shag the stewardesses at the destination?

I think we will see the FedEx type operations move there first. Charter freight to more remote areas. Then small aircraft and personal planes.

On commercial aircraft, there are still technology advances that will come into the cockpit before the pilots leave.

The bigger challenge is how do we keep qualified pilots operating our commercial planes, in the US and other countries, the skill comes with training and experience, we all know there is a time coming where the supply of pilots doesn't match demand. Airlines will need to pay up, support flying programs etc. As the military adopts single pilot and autonomous aircraft, that feeder pool will continue to shrink.

Boof02671 wrote:

How will a computer handle an electrical failure? A bomb?

If this then that. The way all software works.

Sure...then you'll have Russian hackers taking over aircraft.

US POTUS has clearly said he won't allow more automation in aircraft. This thread is pointless.

The self driving cars that we have are more like autopilot. Tesla won't drive at all if you start sleeping on driver's seat while leaving the car on self drive mode.

airzona11 wrote:

I think we will see the FedEx type operations move there first. Charter freight to more remote areas. Then small aircraft and personal planes

While some use their small , personal aircraft for transport... most of us make the traveling part an excuse to go fly!!! gonna have to go all Heston on this one and say "you can take the yoke out of my cold dead hand! (Only 1 because 2 hands on the yoke causes over control ...

airzona11 wrote:

The bigger challenge is how do we keep qualified pilots operating our commercial planes, in the US and other countries, the skill comes with training and experience, we all know there is a time coming where the supply of pilots doesn't match demand. Airlines will need to pay up, support flying programs etc. As the military adopts single pilot and autonomous .

The answer is getting kids into aviation , and wit. Airports so restricted (MSP a very awesome exception with their viewing area)kids don't see airplanes and don't get interested ..
then again with all this automation makes everything better BS , who'd want to spend 100k to get replaced by a robot ?

I enjoy reading these threads. Not because I think it's an interesting topic, but because of the misinformation of what the automation on a modern plane is actually capable of doing. Take for example the aircraft I fly, the A320 family of planes. Let me take you on a leg of my day of what I have to do that automation cannot. It's winter time so I'll just use a winter day as my example. For this one, we'll go from ORD-IAH and assume there's a weather system between here and there.

I arrive in the cockpit and notice already there's a nuisance ECAM. I know this message will clear, its common that it pops up sometimes when a lot of fuel is being put on the aircraft. I watch it but I know it'll clear in a minute or two. If this plane was self reporting, maybe it's going to summon maintenance unnecessarily causing wasted time and delays. How would the plane know? How will maintenance know? How will maintenance know the plane was being fueled triggering the nuisance. Anyways small potatoes, let me go do my walk around.

I walk around and see fluid dripping from the engine cowling. Our aircraft is a NEO, any fluid dripping from the cowling is a no go item. I look closer, its residual deicing fluid that has run from the wing down the cowling. I bring this example up because I actually found a small fluid drip on a NEO that fueled out to be fuel. Ended up cancelling a flight because it revealed a lot broken fuel system parts after the cowling was open. Inside the aircraft? There were no indications of any issues. We took about a hour delay because the broken equipment was discovered. This job can easily be solved by having a mechanic do a walk around but lets say I'm at an outstation that utilizes contract maintenance? Who is going to figure out this is a problem or just water or deicing fluid? Beats me. The airplane sure isn't going to.

I get back in the plane, we're done fueling. It's cold out, about a degree above freezing, but there's no falling precip. Calm winds. On my walk around the leading edges of the wings, stabs were all clean of ice. I walk back to the exit rows and look out the windows. Sure enough the top of the wings over the fuel tanks have frosted up from the much cooler fuel. We have to deice. The plane isn't going to know there's ice on this part of the wing. Only human eyes can see this. We call for deicing after pushback.

We call for push. There are no planes in the alley, the push goes as planned until the towbar breaks. Crappy towbar had worn sheer pins and now the aircraft is no longer attached to the tug and is rolling backwards towards parked airplanes. The headset cord from the tug to the plane breaks and there no longer is communication to the cockpit. We see the tug appear further than typical from the plane and rampers are frantically waving their arms at us. One of us steps on the brakes and stops. Crisis averted. Would the plane have known on it's own it is freely rolling backwards towards disaster? Lets say the push goes as planned, who commands to start the engines of the aircraft? Short taxi both engines are started but what if there's a money line for 28L at N5 at ORD. 20 planes deep, how is the aircraft going to know it won't be taking off for a half hour or so? Will it just burn gas for no reason? We get our taxi instructions. A A15 N to 28R/N5. We enter taxiway A and on one of the cross roads, a tug brakes late and ends up about 4-5 feet past the stop line. We going to clear the tug by about 40' but it's across the safety line. Do we stop or do we continue? Human intervention strikes again.

The money line is too deep and A15 taxiway isn't going to work. We get told by Ohare ground to continue on A all the way down to A9 and now T and N and to hold for two RJ's from the right who will join in sequence ahead of us but a third Ejet will give way for us. Who inputs this information into the aircraft if it had an auto taxi function? Long taxi the brake temperature is slowly climbing. At 300 degrees, we have to pull off to the side until they cool, take off is prohibited. We gotta really baby the brakes. It's cold out but it doesn't matter. The brake temps slowly approach 200. Due to some snow that hasn't cleared yet on the taxiways, we turn on the engine anti ice. We won't need it for take off but we can't risk ingesting ice on the taxiway. We wait in line, now 10 to go, and it starts snowing. Light snow but it's still snow. Lake Michigan is a tricky SOB. When we deiced we got type 1 and IV deicing fluid. We're well within our holdover time but now the runway is contaminated. The ATIS isn't calling snow yet but we see it with our eyes. We decide that we need to re run performance numbers with engine anti ice on. A computer would never be able to make this determination.

We see we're about five minutes for take off, we start the second engine. Get to the end of the runway and are cleared for take off. We're given the O'HARE 5 departure and a turn to head 180. At 5.5 DME from the Chicago VOR, we need to be at or above 3000' MSL. A wind from the north has really got our ground speed clipping. We're not going to make the restriction unless we continue with our take off power and a hold off on trying to get to 250kts. The flight director is telling me to lower the nose to accelerate and FMA is telling me to go to climb power. I'm still within MCT/FLEX time limits on the plane so I continue to climb for a few more seconds in this configuration just to make sure we make the first restriction. If I turned the AP on as soon as I could and followed all aircraft instructions, not a chance I come close to complying with ATC.

This post is getting long so I'll just skip ahead to cruise.Flight level 340, AP on (hey look the airplane is "doing" all the flying). Around Little Rock, we see the beginning of a weather system that's going to follow us all the way to Houston. While we try to determine a game plane to pick our way around build ups, ATC asks us for crossing traffic, can we go up or down to an odd altitude. Looking at the weather ahead, we determine going up would get us clear of a majority of the clouds. The winds are stronger up there but down low we're going to have to take some vectors to get around some buildups we can see about 100 miles ahead. The radar isn't painting anything but we can see we can easily get jammed into a corner down low. As we get closer to the weather, some of the developing storms are well over the service ceiling of the plane. The radar is working but it's not working great. A few cells are causing attenuation. We want to stay away from those. There's clearing between two cells to our east on the radar but we can see with our eyes the giant overhang that fills the ceiling. To our west, there's a gap but it's a bit of a further distance away. If we can get there, we will be west of the weather that is just getting to Houston.

Continuing on our west vectors, things are getting bumpy but not too bad. It light turbulence but we turn on the seat belt sign. The turbulence gets stronger and we're experiencing occasional moderate. ATC informs us it's a better ride at FL300. We crunch the numbers with fuel and see that we can do it. East arrivals into IAH have been closed, the line of weather is getting heavy and aircraft are starting to hold. We're on the DRLLR.5 arrival. It's our lucky day. We're leading the pack and told to descend VIA the DRLLR.5 except maintain 300kts or greater. The DRLLR.5 has all sorts of speed restrictions on it and we don't need to follow any of them until we get past 10000'. The A320 has VNAV capabilities, the AP is working great but because of the amendment to the speed restrictions, I have the speed selected at 310kts. We want to get into Houston ASAP because the weather is a big question mark. We're screeming down when ATC comes back to us and tells us to resume published speeds at the DRLLR fix. The published speed is 250kts. The average rule of thumb in level flight is 10kts of speed can be lost for every 1NM. We're still on a descent and thanks to airbus logic,it always aims for the bottom of any altitude restrictions. We're going to cross DRLLR at 13000 even though it's published to be crossed between 16000-13000' but at the selected speed of 310kts. I need now to figure out how to get off the descent profile and lose the speed. If I cross the fix higher than the ideal aircraft profile, the airbus will snowball the high altitude and blow the rest of the arrival without human intervention. I set the speed for 320kts and pull for an open descent and deploy as much speedbrakes as I can. We'll level off at 13000' about 3-4 miles before DRLLR and miss the speed by maybe 10kts but it's close enough. ATC is happy and we continue on the descent path. Never in a million years would the automation on the aircraft been able to make a decision that'll take it off the profile to make a new restriction.

We're landing on runway 26R. It's still raining but the meat of the line of weather is 20 miles to the east of the airport. Winds are gusting up to 25kts from the north giving us a pretty good crosswind. ATC has us at 4000' feet MSL headed east. We're on radar vectors now. We fly past the airport. We're now 10 miles past the airport. We can see the back of the line of storms that just pushed through. We need to turn and we need to turn now back towards the airport. Using strategic language and a tone usually reserved for complaining about bad food at a restaurant, we get the approach controllers attention. We get a turn to the south and an eventual turn to intercept the final approach course. We're told to hold 210 kts. We have flaps 1 out but that's all we can do to slow other that dropping the gear. We're cleared for the ILS26R. Can't use an autoland (magic autoland) because of the crosswind and wet runway. We glide slope intercept at RAIDS fix on the approach and are still at 210kts. We need to slow and now we're fast, in gusts, and headed downhill. I ask for the gear, I pull the speedbrakes again, we're only slowing to 200. Not good enough for flaps 2. I click the auto pilot off and fly above the glide slow to shallow the descent so lose speed. We're one dot high and but enough for flaps two, it's working and we're able to get fully configured for landing. Break out of the clouds at 1500', stable at 1000'. Manually land with a 20 something knot crosswind and taxi off at the nearest high speed turn off.

Houston is full of taxiway construction, we're asked due to the mess the weather has caused to taxi all the way around to the east side of the terminal to go around the backside into the alley to find our gates in Terminal A. 40 minutes later through the congestion, we get to the gate. It's occupied with a flight that's delayed. We need to coordinate with the ramp to park our plane somewhere that's not in the way. Operations is telling us how much time we have until we have a gate. We finally park. We pull into the gate. We run a parking checklist. There's a ding. We got a nuisance ECAM. We still have to do one more flight out east.

This is a daily occurrence for pilots all over the world. I want someone to tell me how a self flying plane would handle this. This is how far we are from pilotless planes.

Do not underestimate progress in computer science, but also do not overestimate progress in aviation industry.
In 20 years, we will have 787 Max, iteration of current tech. In 50 years, we will have still the same tube with wings, still a 797 derivates.

If-else programming cannot do any problem solving. Self learning can solve some problems, but will be very unpridictable. Sometimes very clever, sometimes extremely stupid. Self aware systems will have own problems, including motivation. Why it must serve humans and fly planes if there are more interesting things to do?

SurlyBonds wrote:

SierraPacific wrote:

is foolish to think that we are even within 50 years of this happening....The only way that a plane could operate autonomously would be with a full fledged AI that can receive data and coordinate on the same level as a human person. The worst part about this is that every person would be hitting the unemployment line at the same time as pilots.

Ever heard of Moore's law?

Ray Kurzweil says that computers may surpass human level intelligence by 2029. Masayoshi Son puts the date at 2045. So some of the best minds in AI disagree with you.

I hate to say it but it seems like some of the posts here are why industry professsionals have been leaving the site at an alarming rate.

Aviation industry professionals are less likely to be knowledgeable about AI than computer scientists.

and if the singularity happens then so be it because a self-aware AI is what it will take to fly an airplane from a to b with all of the wrenches professional pilots deal with on a daily basis. I would bet money that you will have surgery done by a robot before you step on a pilotless plane.

n6238p wrote:

I enjoy reading these threads. Not because I think it's an interesting topic, but because of the misinformation of what the automation on a modern plane is actually capable of doing. Take for example the aircraft I fly, the A320 family of planes. Let me take you on a leg of my day of what I have to do that automation cannot. It's winter time so I'll just use a winter day as my example. For this one, we'll go from ORD-IAH and assume there's a weather system between here and there.

I arrive in the cockpit and notice already there's a nuisance ECAM. I know this message will clear, its common that it pops up sometimes when a lot of fuel is being put on the aircraft. I watch it but I know it'll clear in a minute or two. If this plane was self reporting, maybe it's going to summon maintenance unnecessarily causing wasted time and delays. How would the plane know? How will maintenance know? How will maintenance know the plane was being fueled triggering the nuisance. Anyways small potatoes, let me go do my walk around.

I walk around and see fluid dripping from the engine cowling. Our aircraft is a NEO, any fluid dripping from the cowling is a no go item. I look closer, its residual deicing fluid that has run from the wing down the cowling. I bring this example up because I actually found a small fluid drip on a NEO that fueled out to be fuel. Ended up cancelling a flight because it revealed a lot broken fuel system parts after the cowling was open. Inside the aircraft? There were no indications of any issues. We took about a hour delay because the broken equipment was discovered. This job can easily be solved by having a mechanic do a walk around but lets say I'm at an outstation that utilizes contract maintenance? Who is going to figure out this is a problem or just water or deicing fluid? Beats me. The airplane sure isn't going to.

I get back in the plane, we're done fueling. It's cold out, about a degree above freezing, but there's no falling precip. Calm winds. On my walk around the leading edges of the wings, stabs were all clean of ice. I walk back to the exit rows and look out the windows. Sure enough the top of the wings over the fuel tanks have frosted up from the much cooler fuel. We have to deice. The plane isn't going to know there's ice on this part of the wing. Only human eyes can see this. We call for deicing after pushback.

We call for push. There are no planes in the alley, the push goes as planned until the towbar breaks. Crappy towbar had worn sheer pins and now the aircraft is no longer attached to the tug and is rolling backwards towards parked airplanes. The headset cord from the tug to the plane breaks and there no longer is communication to the cockpit. We see the tug appear further than typical from the plane and rampers are frantically waving their arms at us. One of us steps on the brakes and stops. Crisis averted. Would the plane have known on it's own it is freely rolling backwards towards disaster? Lets say the push goes as planned, who commands to start the engines of the aircraft? Short taxi both engines are started but what if there's a money line for 28L at N5 at ORD. 20 planes deep, how is the aircraft going to know it won't be taking off for a half hour or so? Will it just burn gas for no reason? We get our taxi instructions. A A15 N to 28R/N5. We enter taxiway A and on one of the cross roads, a tug brakes late and ends up about 4-5 feet past the stop line. We going to clear the tug by about 40' but it's across the safety line. Do we stop or do we continue? Human intervention strikes again.

The money line is too deep and A15 taxiway isn't going to work. We get told by Ohare ground to continue on A all the way down to A9 and now T and N and to hold for two RJ's from the right who will join in sequence ahead of us but a third Ejet will give way for us. Who inputs this information into the aircraft if it had an auto taxi function? Long taxi the brake temperature is slowly climbing. At 300 degrees, we have to pull off to the side until they cool, take off is prohibited. We gotta really baby the brakes. It's cold out but it doesn't matter. The brake temps slowly approach 200. Due to some snow that hasn't cleared yet on the taxiways, we turn on the engine anti ice. We won't need it for take off but we can't risk ingesting ice on the taxiway. We wait in line, now 10 to go, and it starts snowing. Light snow but it's still snow. Lake Michigan is a tricky SOB. When we deiced we got type 1 and IV deicing fluid. We're well within our holdover time but now the runway is contaminated. The ATIS isn't calling snow yet but we see it with our eyes. We decide that we need to re run performance numbers with engine anti ice on. A computer would never be able to make this determination.

We see we're about five minutes for take off, we start the second engine. Get to the end of the runway and are cleared for take off. We're given the O'HARE 5 departure and a turn to head 180. At 5.5 DME from the Chicago VOR, we need to be at or above 3000' MSL. A wind from the north has really got our ground speed clipping. We're not going to make the restriction unless we continue with our take off power and a hold off on trying to get to 250kts. The flight director is telling me to lower the nose to accelerate and FMA is telling me to go to climb power. I'm still within MCT/FLEX time limits on the plane so I continue to climb for a few more seconds in this configuration just to make sure we make the first restriction. If I turned the AP on as soon as I could and followed all aircraft instructions, not a chance I come close to complying with ATC.

This post is getting long so I'll just skip ahead to cruise.Flight level 340, AP on (hey look the airplane is "doing" all the flying). Around Little Rock, we see the beginning of a weather system that's going to follow us all the way to Houston. While we try to determine a game plane to pick our way around build ups, ATC asks us for crossing traffic, can we go up or down to an odd altitude. Looking at the weather ahead, we determine going up would get us clear of a majority of the clouds. The winds are stronger up there but down low we're going to have to take some vectors to get around some buildups we can see about 100 miles ahead. The radar isn't painting anything but we can see we can easily get jammed into a corner down low. As we get closer to the weather, some of the developing storms are well over the service ceiling of the plane. The radar is working but it's not working great. A few cells are causing attenuation. We want to stay away from those. There's clearing between two cells to our east on the radar but we can see with our eyes the giant overhang that fills the ceiling. To our west, there's a gap but it's a bit of a further distance away. If we can get there, we will be west of the weather that is just getting to Houston.

Continuing on our west vectors, things are getting bumpy but not too bad. It light turbulence but we turn on the seat belt sign. The turbulence gets stronger and we're experiencing occasional moderate. ATC informs us it's a better ride at FL300. We crunch the numbers with fuel and see that we can do it. East arrivals into IAH have been closed, the line of weather is getting heavy and aircraft are starting to hold. We're on the DRLLR.5 arrival. It's our lucky day. We're leading the pack and told to descend VIA the DRLLR.5 except maintain 300kts or greater. The DRLLR.5 has all sorts of speed restrictions on it and we don't need to follow any of them until we get past 10000'. The A320 has VNAV capabilities, the AP is working great but because of the amendment to the speed restrictions, I have the speed selected at 310kts. We want to get into Houston ASAP because the weather is a big question mark. We're screeming down when ATC comes back to us and tells us to resume published speeds at the DRLLR fix. The published speed is 250kts. The average rule of thumb in level flight is 10kts of speed can be lost for every 1NM. We're still on a descent and thanks to airbus logic,it always aims for the bottom of any altitude restrictions. We're going to cross DRLLR at 13000 even though it's published to be crossed between 16000-13000' but at the selected speed of 310kts. I need now to figure out how to get off the descent profile and lose the speed. If I cross the fix higher than the ideal aircraft profile, the airbus will snowball the high altitude and blow the rest of the arrival without human intervention. I set the speed for 320kts and pull for an open descent and deploy as much speedbrakes as I can. We'll level off at 13000' about 3-4 miles before DRLLR and miss the speed by maybe 10kts but it's close enough. ATC is happy and we continue on the descent path. Never in a million years would the automation on the aircraft been able to make a decision that'll take it off the profile to make a new restriction.

We're landing on runway 26R. It's still raining but the meat of the line of weather is 20 miles to the east of the airport. Winds are gusting up to 25kts from the north giving us a pretty good crosswind. ATC has us at 4000' feet MSL headed east. We're on radar vectors now. We fly past the airport. We're now 10 miles past the airport. We can see the back of the line of storms that just pushed through. We need to turn and we need to turn now back towards the airport. Using strategic language and a tone usually reserved for complaining about bad food at a restaurant, we get the approach controllers attention. We get a turn to the south and an eventual turn to intercept the final approach course. We're told to hold 210 kts. We have flaps 1 out but that's all we can do to slow other that dropping the gear. We're cleared for the ILS26R. Can't use an autoland (magic autoland) because of the crosswind and wet runway. We glide slope intercept at RAIDS fix on the approach and are still at 210kts. We need to slow and now we're fast, in gusts, and headed downhill. I ask for the gear, I pull the speedbrakes again, we're only slowing to 200. Not good enough for flaps 2. I click the auto pilot off and fly above the glide slow to shallow the descent so lose speed. We're one dot high and but enough for flaps two, it's working and we're able to get fully configured for landing. Break out of the clouds at 1500', stable at 1000'. Manually land with a 20 something knot crosswind and taxi off at the nearest high speed turn off.

Houston is full of taxiway construction, we're asked due to the mess the weather has caused to taxi all the way around to the east side of the terminal to go around the backside into the alley to find our gates in Terminal A. 40 minutes later through the congestion, we get to the gate. It's occupied with a flight that's delayed. We need to coordinate with the ramp to park our plane somewhere that's not in the way. Operations is telling us how much time we have until we have a gate. We finally park. We pull into the gate. We run a parking checklist. There's a ding. We got a nuisance ECAM. We still have to do one more flight out east.

This is a daily occurrence for pilots all over the world. I want someone to tell me how a self flying plane would handle this. This is how far we are from pilotless planes.

Thanks for posting this write up. I think most people, on an aviation forum no less, don't have any idea how aircraft actually are flown at the professional level.

n6238p, that was a nice read, thanks. Another of a million points I have't seen mentioned is a medical emergency. How will a self flying plane know about a passenger medical emergency? Going to leave it to the passengers to hit the big red emergency button and make a costly diversion? Technology is wonderful and has it's places, but self flying planes isn't one of them. I would be willing to say we will never see it in commercial aviation.

In fact, I would go on to say that another form of high speed transportation is more likely to replace the airplane as we know it before we ever see self flying passenger airplanes.

That's a great idea! Now, we also need to get that speedily rolled out, so the AI pilot will only get input from one sensor. And since this is such a minor change anyway, our certification from the 1950s will still be usable.

buzzard302 wrote:

n6238p, that was a nice read, thanks. Another of a million points I have't seen mentioned is a medical emergency. How will a self flying plane know about a passenger medical emergency? Going to leave it to the passengers to hit the big red emergency button and make a costly diversion?

Fully automated metro-trains have had exactly such a system for years. Heck, AFAIK nearly all trains have had some kind of passenger-operated emergency brake since decades ago. All you need to do is put a big warning about a big fine for mis-use on it, put it behind some kind of "break glass" lock and make it trigger communication with an operations centre who can discuss/assess the situation while the aircraft is already automatically diverting. Then operations can decide to direct the aircraft or re-set if necessary.

SomebodyInTLS wrote:

buzzard302 wrote:

n6238p, that was a nice read, thanks. Another of a million points I have't seen mentioned is a medical emergency. How will a self flying plane know about a passenger medical emergency? Going to leave it to the passengers to hit the big red emergency button and make a costly diversion?

Fully automated metro-trains have had exactly such a system for years. Heck, AFAIK nearly all trains have had some kind of passenger-operated emergency brake since decades ago. All you need to do is put a big warning about a big fine for mis-use on it, put it behind some kind of "break glass" lock and make it trigger communication with an operations centre who can discuss/assess the situation while the aircraft is already automatically diverting. Then operations can decide to direct the aircraft or re-set if necessary.

I agree, I am familiar with e-stop buttons in trains and other industrial settings. That's all land based stuff with less overall consequence of an unexpected stop. If you have to have an operations center to real time monitor all these self flying planes, why not just put a pilot on the plane in the first place? Operations centers aren't free, and a decision about what may be occurring on an aircraft 36,000 ft in the air probably shouldn't be made remotely. I'm just a pessimist by nature and tend to see all the negative stuff that "could" happen. Like I said earlier, I could maybe envision a single pilot that takes on more of a monitoring role. Still, it happens all too often that a pilot becomes incapacitated and the second pilot has to finish the flight alone. There are too many logical flaws in having a completely unmanned airplane.

n6238p wrote:

I enjoy reading these threads. Not because I think it's an interesting topic, but because of the misinformation of what the automation on a modern plane is actually capable of doing. Take for example the aircraft I fly, the A320 family of planes. Let me take you on a leg of my day of what I have to do that automation cannot. It's winter time so I'll just use a winter day as my example. For this one, we'll go from ORD-IAH and assume there's a weather system between here and there.

I arrive in the cockpit and notice already there's a nuisance ECAM. I know this message will clear, its common that it pops up sometimes when a lot of fuel is being put on the aircraft. I watch it but I know it'll clear in a minute or two. If this plane was self reporting, maybe it's going to summon maintenance unnecessarily causing wasted time and delays. How would the plane know? How will maintenance know? How will maintenance know the plane was being fueled triggering the nuisance. Anyways small potatoes, let me go do my walk around.

I walk around and see fluid dripping from the engine cowling. Our aircraft is a NEO, any fluid dripping from the cowling is a no go item. I look closer, its residual deicing fluid that has run from the wing down the cowling. I bring this example up because I actually found a small fluid drip on a NEO that fueled out to be fuel. Ended up cancelling a flight because it revealed a lot broken fuel system parts after the cowling was open. Inside the aircraft? There were no indications of any issues. We took about a hour delay because the broken equipment was discovered. This job can easily be solved by having a mechanic do a walk around but lets say I'm at an outstation that utilizes contract maintenance? Who is going to figure out this is a problem or just water or deicing fluid? Beats me. The airplane sure isn't going to.

I get back in the plane, we're done fueling. It's cold out, about a degree above freezing, but there's no falling precip. Calm winds. On my walk around the leading edges of the wings, stabs were all clean of ice. I walk back to the exit rows and look out the windows. Sure enough the top of the wings over the fuel tanks have frosted up from the much cooler fuel. We have to deice. The plane isn't going to know there's ice on this part of the wing. Only human eyes can see this. We call for deicing after pushback.

We call for push. There are no planes in the alley, the push goes as planned until the towbar breaks. Crappy towbar had worn sheer pins and now the aircraft is no longer attached to the tug and is rolling backwards towards parked airplanes. The headset cord from the tug to the plane breaks and there no longer is communication to the cockpit. We see the tug appear further than typical from the plane and rampers are frantically waving their arms at us. One of us steps on the brakes and stops. Crisis averted. Would the plane have known on it's own it is freely rolling backwards towards disaster? Lets say the push goes as planned, who commands to start the engines of the aircraft? Short taxi both engines are started but what if there's a money line for 28L at N5 at ORD. 20 planes deep, how is the aircraft going to know it won't be taking off for a half hour or so? Will it just burn gas for no reason? We get our taxi instructions. A A15 N to 28R/N5. We enter taxiway A and on one of the cross roads, a tug brakes late and ends up about 4-5 feet past the stop line. We going to clear the tug by about 40' but it's across the safety line. Do we stop or do we continue? Human intervention strikes again.

The money line is too deep and A15 taxiway isn't going to work. We get told by Ohare ground to continue on A all the way down to A9 and now T and N and to hold for two RJ's from the right who will join in sequence ahead of us but a third Ejet will give way for us. Who inputs this information into the aircraft if it had an auto taxi function? Long taxi the brake temperature is slowly climbing. At 300 degrees, we have to pull off to the side until they cool, take off is prohibited. We gotta really baby the brakes. It's cold out but it doesn't matter. The brake temps slowly approach 200. Due to some snow that hasn't cleared yet on the taxiways, we turn on the engine anti ice. We won't need it for take off but we can't risk ingesting ice on the taxiway. We wait in line, now 10 to go, and it starts snowing. Light snow but it's still snow. Lake Michigan is a tricky SOB. When we deiced we got type 1 and IV deicing fluid. We're well within our holdover time but now the runway is contaminated. The ATIS isn't calling snow yet but we see it with our eyes. We decide that we need to re run performance numbers with engine anti ice on. A computer would never be able to make this determination.

We see we're about five minutes for take off, we start the second engine. Get to the end of the runway and are cleared for take off. We're given the O'HARE 5 departure and a turn to head 180. At 5.5 DME from the Chicago VOR, we need to be at or above 3000' MSL. A wind from the north has really got our ground speed clipping. We're not going to make the restriction unless we continue with our take off power and a hold off on trying to get to 250kts. The flight director is telling me to lower the nose to accelerate and FMA is telling me to go to climb power. I'm still within MCT/FLEX time limits on the plane so I continue to climb for a few more seconds in this configuration just to make sure we make the first restriction. If I turned the AP on as soon as I could and followed all aircraft instructions, not a chance I come close to complying with ATC.

This post is getting long so I'll just skip ahead to cruise.Flight level 340, AP on (hey look the airplane is "doing" all the flying). Around Little Rock, we see the beginning of a weather system that's going to follow us all the way to Houston. While we try to determine a game plane to pick our way around build ups, ATC asks us for crossing traffic, can we go up or down to an odd altitude. Looking at the weather ahead, we determine going up would get us clear of a majority of the clouds. The winds are stronger up there but down low we're going to have to take some vectors to get around some buildups we can see about 100 miles ahead. The radar isn't painting anything but we can see we can easily get jammed into a corner down low. As we get closer to the weather, some of the developing storms are well over the service ceiling of the plane. The radar is working but it's not working great. A few cells are causing attenuation. We want to stay away from those. There's clearing between two cells to our east on the radar but we can see with our eyes the giant overhang that fills the ceiling. To our west, there's a gap but it's a bit of a further distance away. If we can get there, we will be west of the weather that is just getting to Houston.

Continuing on our west vectors, things are getting bumpy but not too bad. It light turbulence but we turn on the seat belt sign. The turbulence gets stronger and we're experiencing occasional moderate. ATC informs us it's a better ride at FL300. We crunch the numbers with fuel and see that we can do it. East arrivals into IAH have been closed, the line of weather is getting heavy and aircraft are starting to hold. We're on the DRLLR.5 arrival. It's our lucky day. We're leading the pack and told to descend VIA the DRLLR.5 except maintain 300kts or greater. The DRLLR.5 has all sorts of speed restrictions on it and we don't need to follow any of them until we get past 10000'. The A320 has VNAV capabilities, the AP is working great but because of the amendment to the speed restrictions, I have the speed selected at 310kts. We want to get into Houston ASAP because the weather is a big question mark. We're screeming down when ATC comes back to us and tells us to resume published speeds at the DRLLR fix. The published speed is 250kts. The average rule of thumb in level flight is 10kts of speed can be lost for every 1NM. We're still on a descent and thanks to airbus logic,it always aims for the bottom of any altitude restrictions. We're going to cross DRLLR at 13000 even though it's published to be crossed between 16000-13000' but at the selected speed of 310kts. I need now to figure out how to get off the descent profile and lose the speed. If I cross the fix higher than the ideal aircraft profile, the airbus will snowball the high altitude and blow the rest of the arrival without human intervention. I set the speed for 320kts and pull for an open descent and deploy as much speedbrakes as I can. We'll level off at 13000' about 3-4 miles before DRLLR and miss the speed by maybe 10kts but it's close enough. ATC is happy and we continue on the descent path. Never in a million years would the automation on the aircraft been able to make a decision that'll take it off the profile to make a new restriction.

We're landing on runway 26R. It's still raining but the meat of the line of weather is 20 miles to the east of the airport. Winds are gusting up to 25kts from the north giving us a pretty good crosswind. ATC has us at 4000' feet MSL headed east. We're on radar vectors now. We fly past the airport. We're now 10 miles past the airport. We can see the back of the line of storms that just pushed through. We need to turn and we need to turn now back towards the airport. Using strategic language and a tone usually reserved for complaining about bad food at a restaurant, we get the approach controllers attention. We get a turn to the south and an eventual turn to intercept the final approach course. We're told to hold 210 kts. We have flaps 1 out but that's all we can do to slow other that dropping the gear. We're cleared for the ILS26R. Can't use an autoland (magic autoland) because of the crosswind and wet runway. We glide slope intercept at RAIDS fix on the approach and are still at 210kts. We need to slow and now we're fast, in gusts, and headed downhill. I ask for the gear, I pull the speedbrakes again, we're only slowing to 200. Not good enough for flaps 2. I click the auto pilot off and fly above the glide slow to shallow the descent so lose speed. We're one dot high and but enough for flaps two, it's working and we're able to get fully configured for landing. Break out of the clouds at 1500', stable at 1000'. Manually land with a 20 something knot crosswind and taxi off at the nearest high speed turn off.

Houston is full of taxiway construction, we're asked due to the mess the weather has caused to taxi all the way around to the east side of the terminal to go around the backside into the alley to find our gates in Terminal A. 40 minutes later through the congestion, we get to the gate. It's occupied with a flight that's delayed. We need to coordinate with the ramp to park our plane somewhere that's not in the way. Operations is telling us how much time we have until we have a gate. We finally park. We pull into the gate. We run a parking checklist. There's a ding. We got a nuisance ECAM. We still have to do one more flight out east.

This is a daily occurrence for pilots all over the world. I want someone to tell me how a self flying plane would handle this. This is how far we are from pilotless planes.

Thank you for that fun and informative story!

buzzard302 wrote:

If you have to have an operations center to real time monitor all these self flying planes, why not just put a pilot on the plane in the first place? Operations centers aren't free, and a decision about what may be occurring on an aircraft 36,000 ft in the air probably shouldn't be made remotely.

You're seriously thinking a global operations centre (possibly multi-airline) staffed by a few dozen people is as expensive than thousands of pilots flying plus thousands more on the ground at any time...?!

And you're also seriously questioning the decision making by staff trained and dedicated to dealing with emergency response? Who do you think would be more likely to correctly risk-assess a medical situation - a specialist operator with access to all sorts of additional information, trained medics on hand, communication with nearest medical facilities on the ground... or your average captain?

FermiParadox wrote:

It's not 50 years away. At all. It's 20, at most.

seriously, you are so far off the mark, its funny.
the raw technology might be there in 20 (some would argue its there now) - but the business, safety and social aspects absolutely will not.
the only thing you may be close on is the downgauge from 2 pilots aboard to 1.

20 years. Who could say that with a straight face? In 20 years we'll have the 797, which Boeing has not indicated will have any sort of dramatic increase in automation. Same with the 777X. If the 797 won't have it, then the NMA won't since it's been described as being similar to the 757/767 relationship. There's no way that they would change that much on the 787X. Same with the A350NEO. The A320 replacement is a bit of a wild card, but Airbus has indicated that it will incorporate much of the technologies seen on the A220, so it's unlikely that'll feature too many changes. The A220NEO won't have too many drastic changes. So, already, the 2 companies in 20 years won't be planning to implement anything drastic, and many of these aircraft will be continuing on for another 15-20 years. The 777X replacement is the only aircraft that could potentially change things, and if it doesn't, then we're around 30-40 years before manufacturers even begin to experiment with this. It may happen, but none of us pilots will see it during our careers, that's for sure, and personally, I wouldn't mind that.

I believe it will happen. But within my lifetime? Not so sure. Airplanes will need to be built from scratch for this role. I am looking at 40-50 years tops. Now many more generations will I see? 2-3? They also need to create some form of thinking AI that can react not from a pre-defined algorithm. But by using it or a collective experience when bad things happen.

ferren wrote:

Self aware systems will have own problems

Has anybody successfully created a truly self aware system? I only heard that AI can solve problems they never encountered before, but how a self aware system works is something beyond my comprehension.

triple3driver wrote:

Just yesterday, we were on final approach to Kennedy, and we had to go around last minute due to a large flock of birds that had decided to come by and say hi. That same day, we were over Britain, and it started to get a little bumpy, and I preemptively decided to turn on the seatbelt signs and requested a decent, which was approved. Then, chaos erupted above us, and luckily we only got it for a little bit. All in one flight. And these aren't uncommon, this is stuff, along with much more, that pilots have to deal with. Good luck getting a computer to do all that, having to deal with these idiots flying drones near airports, birds, making the decision to go around, extreme turbulence, collision avoidance, and much more, while dealing with everything else in addition to emergencies. I've been flying for 29 years, and for 29 years I've been hearing about computers replacing humans in the near future, and I'll keep hearing it for another 29 years.

You have been flying for 29 years?

In a post I just quoted, you said you were in your mid-40’s and with Delta Air Lines (not Northwest or other predecessor companies) for 17 years and an A330 Captain. Even with the “fencing” during the integration, the absence of said fencing post-merger, this doesn’t make any sense seniority-wise/age calculation.

Which base are you?

triple3driver wrote:

Just yesterday, we were on final approach to Kennedy, and we had to go around last minute due to a large flock of birds that had decided to come by and say hi. That same day, we were over Britain, and it started to get a little bumpy, and I preemptively decided to turn on the seatbelt signs and requested a decent, which was approved. Then, chaos erupted above us, and luckily we only got it for a little bit. All in one flight. And these aren't uncommon, this is stuff, along with much more, that pilots have to deal with. Good luck getting a computer to do all that, having to deal with these idiots flying drones near airports, birds, making the decision to go around, extreme turbulence, collision avoidance, and much more, while dealing with everything else in addition to emergencies. I've been flying for 29 years, and for 29 years I've been hearing about computers replacing humans in the near future, and I'll keep hearing it for another 29 years.

Where and when on approach to Kennedy did you encounter a large flock of birds? There is no corresponding ASAP report in the system...how about a flight number and date to verify the Go-Around?

You are on the A330 fleet as a Captain. When and how could you have encountered a “large flock of birds” that prompted a go-around? JFK has a permanent ATIS message about birds in and on the airport, and, nothing I can think of in a FOM or in SOP expressly calls for a “go around” due to a “large flock of birds”...they would have to be nesting on the threshold....

The A330 can certainly handle a “bird strike” or a few, and, in my experience, I’ve never had to outmaneuver a “large flock of birds” to the point of a “go around”. Birds are notoriously slow-moving and incredibly hard to spot in the visual scan, especially when flying a landing-configured, on-speed, on glidepath A330 in VMC or IMC. I have never, ever had to say “Let’s go missed, there are birds over the outer marker”....

I think your stories are great, they are just not true.