Britten-Norman Pilotless Planes
Tesla are coming out with an autopilot which will cope with suburban street driving. This is a far more complex, uncontrolled and random environment than aviation.
As for 30 kts of crosswind? The autopilot can easily do it with more precision than any human. Likewise engine failure or anything else you care to throw at it.
Until recently, autopilots didn’t have “eyes”. Not so anymore. Image processing and AI have leapt forward in leaps and bounds. We have moved on from when the Century III was the pinnacle of technology.
Also note that Garmin now have a $250K option for their autopilots on the new Piper Meridian - a large red button which tells the A/P to land at nearest suitable airport in case of the (frail) human pilot becoming disabled. A/P does it all including configuration and selecting gear down and neither does it need a Cat 3 ILS to land the aircraft.
The good old days of steely eyed pilots in leather flying jackets being an essential part of air travel are coming to an end.
Your airliner of the future will be controlled from a repurposed call centre in Mumbai.
As for 30 kts of crosswind? The autopilot can easily do it with more precision than any human. Likewise engine failure or anything else you care to throw at it.
Until recently, autopilots didn’t have “eyes”. Not so anymore. Image processing and AI have leapt forward in leaps and bounds. We have moved on from when the Century III was the pinnacle of technology.
Also note that Garmin now have a $250K option for their autopilots on the new Piper Meridian - a large red button which tells the A/P to land at nearest suitable airport in case of the (frail) human pilot becoming disabled. A/P does it all including configuration and selecting gear down and neither does it need a Cat 3 ILS to land the aircraft.
The good old days of steely eyed pilots in leather flying jackets being an essential part of air travel are coming to an end.
Your airliner of the future will be controlled from a repurposed call centre in Mumbai.
Let's see how a pilotless BN2 can cope with the weather & airstrips in Shetland. I flew there & I just don't see automation technology being anywhere near having the ability, robustness, and adaptabilityin those sorts of places.
Tesla are coming out with an autopilot which will cope with suburban street driving. This is a far more complex, uncontrolled and random environment than aviation.
As for 30 kts of crosswind? The autopilot can easily do it with more precision than any human. Likewise engine failure or anything else you care to throw at it.
As for 30 kts of crosswind? The autopilot can easily do it with more precision than any human. Likewise engine failure or anything else you care to throw at it.
Join Date: Jan 2014
Location: Melbourne
Age: 68
Posts: 3
Likes: 0
Received 0 Likes
on
0 Posts
Yawn.
Who liaises with Operations / Airport staff / Flight Attendants / Ground Engineers / clearance delivery / ATC / Met. Department etc. etc. ?
Oh , that would be a Pilot.
The delusion behind all this is that all a Pilot does is manipulate the aircraft.
The only people enthusiast about this are those that have never operated an RPT jet.
It isn’t happening anytime soon.
Who liaises with Operations / Airport staff / Flight Attendants / Ground Engineers / clearance delivery / ATC / Met. Department etc. etc. ?
Oh , that would be a Pilot.
The delusion behind all this is that all a Pilot does is manipulate the aircraft.
The only people enthusiast about this are those that have never operated an RPT jet.
It isn’t happening anytime soon.
Join Date: Jan 2008
Location: turn L @ Taupo, just past the Niagra Falls...
Posts: 596
Received 0 Likes
on
0 Posts
Originally Posted by White Knight
How things have changed! I managed to fly both Islanders and Trislanders without advice from an ‘autonomous co-pilot’🤔😳...
Me too
this site Seriously needs a Like button!
Avoid imitations
Join Date: Nov 2000
Location: Wandering the FIR and cyberspace often at highly unsociable times
Posts: 14,573
Received 422 Likes
on
222 Posts
This could be just as successful as the Firecracker!
Join Date: Dec 2015
Location: France
Posts: 507
Likes: 0
Received 0 Likes
on
0 Posts
The competencies "procedures" and "decision making" involve : "knows how and when to deviate from standard procedures when safety requires so"
Not being able to do this results in an "unacceptable" mark.
People were kicked out of training at some points because they were unable to deviate from procedures. For example one guy refused to put the gear out before flaps 2 and after flaps 3.
So an autonomous aircraft would also be required to deviate from what it was told, in safety requiring conditions.
Not a pilot, but can I ask why you wouldn't trust automation to carry out correct procedures in any situation? I would have thought that the problems would come in the situations where there are no pre-defined correct procedures, perhaps where the only choice is the lesser of two evils, or in those situations where current automation gives up and hands the aeroplane back to the pilot. Check-lists and SOPs, I gather, are responsible for a lot of the improvement in safety; is there a way in which even SOPs have to be performed with human-only nuance? (A real question.)
Let's just hope that artificial intelligence is not already able to do this.
For a large airline, 20% is the cost of the total staff, including cabin crew, ground crew and office teams..
Pilots may represent only a quarter of that so maximum 5% of the cost of an airline.
Avoid imitations
Join Date: Nov 2000
Location: Wandering the FIR and cyberspace often at highly unsociable times
Posts: 14,573
Received 422 Likes
on
222 Posts
Fully autonomous aircraft will obviously need a backup system.
I can imagine the situation with "George and George" in control:
"You have control....BEEP!"
"Negative....BEEP!"
"BEEP!"
"BEEP!"
Pax: "Eek!"
Well, as ASO said back at Post #2, BAE Systems have been fiddling about(
) with their autonomous Jetstream for some years now. Last I can find on it is https://www.baesystems.com/en/articl...flying-testbed from 2016.
I guess the only advantage of the Islander is that, even in the most sound-proof cockpit, the other crews would detect the Islander coming from miles off without even needing TCAS!
Ironic it is an Islander being used for an autonomous aircraft trial as it's the best fun flying I've ever had! Took off limited fuel on a ferry flight in a "rather stiff headwind" once - which rather confused ATC! They made the silly mistake of looking for me along the normal departure path and had to ask me where I was .......
Went up almost like a Saturn 5 ...................... just a lot noisier than a Saturn 5!!!!
Happy days! H 'n' H
) with their autonomous Jetstream for some years now. Last I can find on it is https://www.baesystems.com/en/articl...flying-testbed from 2016.I guess the only advantage of the Islander is that, even in the most sound-proof cockpit, the other crews would detect the Islander coming from miles off without even needing TCAS!
Ironic it is an Islander being used for an autonomous aircraft trial as it's the best fun flying I've ever had! Took off limited fuel on a ferry flight in a "rather stiff headwind" once - which rather confused ATC! They made the silly mistake of looking for me along the normal departure path and had to ask me where I was .......
Went up almost like a Saturn 5 ...................... just a lot noisier than a Saturn 5!!!!
Happy days! H 'n' H
In practice, I would suggest that 3-axis control law is a precise art and thus autopilots can fly with a precision difficult for humans to equal (consistently anyway). But complex decision making under complex failure conditions is not the same sort of problem to solve. It can of course also be automated, after all that is what emergency checklists are trying to achieve. However, you do need to be sure that you have covered every possible eventuality in the software. You also need to provide the autopilot with actuators for pretty much every lever, button, switch and circuit breaker in the aircraft. And when you accept that then you begin to wonder if retrofit of an existing aircraft design is really an option.
As for the engine fire scenario, just as an example: Shut down and land soonest procedures may be easily automated - but deciding the fire is not contained, structural integrity is at risk and we have to ditch now - that's a much bigger judgement ask for an automatic system.
So basically I do not think we are there yet.
Comparison with Tesla autodrive vehicles is of limited validity, because those autopilots always have the option to simply stop the car when it all gets a bit too much.
....... As for the engine fire scenario, just as an example: Shut down and land soonest procedures may be easily automated - but deciding the fire is not contained, structural integrity is at risk and we have to ditch now - that's a much bigger judgement ask for an automatic system. ......
....... So basically I do not think we are there yet. ......
Would I like it? My main personal reservation is not having someone in control who would arrive at the scene of the accident a few milliseconds before I did as I feel that does help concentrate the minds of those i/c. That said, humans often put themselves into trouble with great panache so that sort of evens the equation up a bit! We've all had those situations when, after a "moment" has happened I've sat back and thought "Mmmm, not your brightest idea that, H 'n' H. Better get on and complete this flight as the pax are sort of relying on you here!".
Will it happen in my lifetime? Military - well, one could argue that we are simply increasing our use of drones which started in 1849 and the use of drones in warfare has now become routine; just another "tool of the trade" of killing (I'm ex-Mil btw). Some cargo? Almost certainly. SLF flights? Probably not in the few decades I may have left ahead of me. But it may be closer than I think. I suspect the biggest hurdle will be the seamless transition from today's airspace environment to an autonomous one.
Interesting times to be sure! H 'n' H
I don't think manipulating the controls is the big problem. I also don't think automating any established normal, abnormal or emergency procedure is the big problem. Neither is configuring the aircraft automatically.
The real challenge is that on most airplanes I have ever flown the abnormal and emergency procedures are far from perfect. On the 737 I thought they were really good in almost(!) all situations I encountered in the sim over the few years I flew the type.
However, most biz jets I have flown there was a lot of room for improvement in the checklists including some obvious mistakes that the manufacturer was very slow to address if at all. I would not want to sit in the back of a plane where a computer is just actioning these kind of imperfect procedures. And that'a before we get into other variables and complexities such as multiple failures, related or unrelated, weather etc.
And what about emergencies such as fire and smoke on board (not engine fires) and medical emergencies? I have had GPS jamming on an aircraft that had no INS/IRS. The DME/DME back up did not work as advertised and the FMS reverted to DR which resulted in a huge map shift in a very short time. How would the automated pilot deal with that?
Yes, the Garmin auto land available on some single pilot GA types today is quite impressive but I am sure it is very far from being able to replace a real pilot day in and day out. It might save the day when the sole pilot of a small plane becomes incapacitated but that's a much lower bar to clear than providing safe transportation in all conditions for billions of passengers on millions of flights every single year.
Not saying it's impossible. The technology is there in principle but it is not nearly ready yet and many years from being implemented widely in my humble opinion. What's needed is true AI which can only work if it has all the inputs (sensors) we have plus the interface in place to interact with it, including input from ATC and mx on the ground, from cabin crew etc.
The real challenge is that on most airplanes I have ever flown the abnormal and emergency procedures are far from perfect. On the 737 I thought they were really good in almost(!) all situations I encountered in the sim over the few years I flew the type.
However, most biz jets I have flown there was a lot of room for improvement in the checklists including some obvious mistakes that the manufacturer was very slow to address if at all. I would not want to sit in the back of a plane where a computer is just actioning these kind of imperfect procedures. And that'a before we get into other variables and complexities such as multiple failures, related or unrelated, weather etc.
And what about emergencies such as fire and smoke on board (not engine fires) and medical emergencies? I have had GPS jamming on an aircraft that had no INS/IRS. The DME/DME back up did not work as advertised and the FMS reverted to DR which resulted in a huge map shift in a very short time. How would the automated pilot deal with that?
Yes, the Garmin auto land available on some single pilot GA types today is quite impressive but I am sure it is very far from being able to replace a real pilot day in and day out. It might save the day when the sole pilot of a small plane becomes incapacitated but that's a much lower bar to clear than providing safe transportation in all conditions for billions of passengers on millions of flights every single year.
Not saying it's impossible. The technology is there in principle but it is not nearly ready yet and many years from being implemented widely in my humble opinion. What's needed is true AI which can only work if it has all the inputs (sensors) we have plus the interface in place to interact with it, including input from ATC and mx on the ground, from cabin crew etc.
Last edited by 733driver; 10th Oct 2020 at 18:18.
Case in point
It has been a while since the pilotless aircraft debate raged on this thread, but the recent KAL incident brought it to mind:
Thankfully no serious injuries were reported. Amid speculation (in the Accidents & Close Calls forum) regarding the precise circumstances leading up to the overrun, I’m surprised there hasn’t been more discussion about how the crew eventually managed to get the plane to stop. At risk of leaping to conclusions based on nothing more than a photograph (see above), the evidence does suggest it went something like this:
If my theory is correct, then it provides a timely reminder of the important role human beings continue to play in the cockpit when unusual and unforeseen circumstances arise. Moreover, I’m not sure a single pair of feet would’ve sufficed on this occasion – this one looks like it was a real ‘four-legger’ (as I suspect this type of close-call will come to be known).
(Despite the tongue-in-cheek observation, there is a genuine point being made here - especially if it turns out that the overrun was related to a fault with the aircraft)
Thankfully no serious injuries were reported. Amid speculation (in the Accidents & Close Calls forum) regarding the precise circumstances leading up to the overrun, I’m surprised there hasn’t been more discussion about how the crew eventually managed to get the plane to stop. At risk of leaping to conclusions based on nothing more than a photograph (see above), the evidence does suggest it went something like this:
If my theory is correct, then it provides a timely reminder of the important role human beings continue to play in the cockpit when unusual and unforeseen circumstances arise. Moreover, I’m not sure a single pair of feet would’ve sufficed on this occasion – this one looks like it was a real ‘four-legger’ (as I suspect this type of close-call will come to be known).
(Despite the tongue-in-cheek observation, there is a genuine point being made here - especially if it turns out that the overrun was related to a fault with the aircraft)
