Pilot Shortage
08-09-2020, 07:51 AM
#21
Gets Weekends Off
Joined APC: May 2013
Posts: 315
how many years do you have building complex software systems? Integrating that with hardware? Doing a security analysis including pen testing? Bluntly, i have way more experience in all those fields than you do, and i can promise you that making the computer "figure it out" is the easiest part of all that.
Proving that it isn't susceptible to black-hat actors is pretty darn hard. Proving it works in a virtually unlimited set of scenarios is harder still. And if there is any "artificial intelligence" involved, proving its functionality approaches the impossible.
In technology, it's often the validation ("testing") rather than the development that tends to drive costs for large scale projects, so this will be an extremely expensive undertaking. Has there been a cost vs. Benefit analysis done to show that such an undertaking is worthwhile?
These all come together to form a pretty complex system. Taking just a thin slice of functionality, you say that the aircraft could be controlled directly from the soc. How are they authenticated (i.e. How does the a/c know it isn't connected to an imposter soc)? How redundant is the soc? Where are they (you weren't thinking it was just a single location, were you?) located? How do they talk to each other? What is the failover protocol? How is a compromised soc detected and locked out? What is the data communications protocol and how much bandwidth will be required?
If you think this all happens quickly, i refer you to ads-b, which is a much simpler system that took decades to materialize. So no, the "head in the sand" folks as you call them, are correct. I have no doubt that such a system will eventually happen, but there is nothing to indicate that it will happen before any current pt. 121 pilot retires.
Proving that it isn't susceptible to black-hat actors is pretty darn hard. Proving it works in a virtually unlimited set of scenarios is harder still. And if there is any "artificial intelligence" involved, proving its functionality approaches the impossible.
In technology, it's often the validation ("testing") rather than the development that tends to drive costs for large scale projects, so this will be an extremely expensive undertaking. Has there been a cost vs. Benefit analysis done to show that such an undertaking is worthwhile?
These all come together to form a pretty complex system. Taking just a thin slice of functionality, you say that the aircraft could be controlled directly from the soc. How are they authenticated (i.e. How does the a/c know it isn't connected to an imposter soc)? How redundant is the soc? Where are they (you weren't thinking it was just a single location, were you?) located? How do they talk to each other? What is the failover protocol? How is a compromised soc detected and locked out? What is the data communications protocol and how much bandwidth will be required?
If you think this all happens quickly, i refer you to ads-b, which is a much simpler system that took decades to materialize. So no, the "head in the sand" folks as you call them, are correct. I have no doubt that such a system will eventually happen, but there is nothing to indicate that it will happen before any current pt. 121 pilot retires.
08-09-2020, 07:54 AM
#22
Gets Weekends Off
Joined APC: Mar 2008
Position: FAA 'Flight Check'
Posts: 13,839
You need my job to enlighten you if you think our systems are perfect.
though there has never been a mishap that I am aware of in the US at least with a causal factor of a false LOC course.
in an case, if the insinuation was that there are “countless” examples of such, that is just sensationalism.
08-09-2020, 07:59 AM
#23
Gets Weekends Off
Joined APC: Mar 2008
Position: FAA 'Flight Check'
Posts: 13,839
No....not a lot, but I am part of the organization who would know if there were even a few cases of false LOCs, and there haven’t been in the last 10 years at least.
If you have seen some - did you report them to ATC.
If you did, the system would immediately be NOTAM’ed off and a SPECIAL flight inspection would be required.
If you have such an example, give me the identifier of the system and I’ll look into it for you.
If you have seen some - did you report them to ATC.
If you did, the system would immediately be NOTAM’ed off and a SPECIAL flight inspection would be required.
If you have such an example, give me the identifier of the system and I’ll look into it for you.
08-09-2020, 08:02 AM
#24
Gets Weekends Off
Joined APC: Dec 2011
Position: A320 FO
Posts: 982
There is more. A lot more. I have issues with the whole text stuff to the airplane path we are on. I build situation awareness listening to the radio. Maybe comes from teaching in SoCal airspace pre tcas. At any rate I can “foresee” issues and problems by listening. I’ve told FO’s This or that’s going to happen. Stay up here a bit. We are following a heavy that they didn’t tell us about etc etc. This ability is lost on automation.
My airplane's crosswind limit is 25 knots. My limit is 45 knots.
It's going to be decades, and will require a clean sheet aircraft and systems architecture design, and a complete change in how the NAS is run. And you've probably seen how quickly the FAA embraces, procures and refines new technology.
Optical sensors are improving? LOL. The USAF can't deliver a FMC refueler based on existing airframe technology after two decades.
Relax. Your concern should be how you're going to keep from having to spend your life savings on medical care. Nearest, hottest threat, buddy.
It's going to be decades, and will require a clean sheet aircraft and systems architecture design, and a complete change in how the NAS is run. And you've probably seen how quickly the FAA embraces, procures and refines new technology.
Optical sensors are improving? LOL. The USAF can't deliver a FMC refueler based on existing airframe technology after two decades.
Relax. Your concern should be how you're going to keep from having to spend your life savings on medical care. Nearest, hottest threat, buddy.
I am in full agreement this isn’t remotely the nearest, hottest threat. I also agree with the decades time frame but will it be two or four? My point is completely dismissing it like many pilots are wont to do is irrational.
08-09-2020, 08:08 AM
#27
Gets Weekends Off
Joined APC: Mar 2008
Position: FAA 'Flight Check'
Posts: 13,839
Identifier of the ILS? Airport? Runway
Did you report it?
What is your definition of an actual FALSE LOC SIGNAL?
The needle moved in what manner.
Tell me the circumstances of the incident.
Are you positive you were within the service volume of the LOC?
Did you report it?
What is your definition of an actual FALSE LOC SIGNAL?
The needle moved in what manner.
Tell me the circumstances of the incident.
Are you positive you were within the service volume of the LOC?
08-09-2020, 08:20 AM
#28
Gets Weekends Off
Joined APC: Apr 2017
Posts: 3,788
Full automation possibly in 50 years, probably in 100 years. But not before then. Programming a plane to fly itself isn't impossible but right now the computers simply can't make decisions with the required reliability.
08-09-2020, 08:31 AM
#29
Gets Weekends Off
Joined APC: Jan 2018
Posts: 644
I also have a hard time buying that airlines will get shrunk down to 1 pilot. With 2 pilots we can turn a jet in 22 minutes on the schedule; take that down to 1 pilot and that'd be at least 45 minutes. 5-6 leg days would never happen. I could maybe see them reducing crew for wide bodies flying over the ocean, at least for the enroute portion, but definitely not domestic flights. Also, pilots are relatively cheap, considering the pay is loosely $1/passenger/hour of flight for FOs and $1.50 per for the CA. I get the narrow profit margins, but to put it in perspective, TSA is $5 per passenger, one-way. After they've lobbied to cut TSA and shown that fully automated planes can operate for a few months without getting violated or over-Ging in towering CB every time they fly into CLT in the summer, then I'll be worried.
08-09-2020, 08:36 AM
#30
Gets Weekends Off
Joined APC: Jan 2018
Posts: 644
Not to mention that after computers have been used all day, every day, for 5 or 6 years, they tend to get "buggy."
Thread
Thread Starter
Forum
Replies
Last Post
