[Efthreeoh]

Originally Posted by JamesNoBrakes

Not really, you don't have to be trained on an autopilot to go buy an aircraft with it and fly it into the ground (or have it installed in your plane). There's no training requirement for that. You appear to maybe be talking about commercial passenger service or something, but even then most of what you said doesn't hold true. I'm talking about private operations, like private car drivers. You should follow the equipment manual, which most people do, but for sure some leave the AP on a little longer and get a little closer than they should. A few just let it take them right into the runwayor the side of a mountain. The only time you have to be "trained" on the autopilot is when you take your check-ride in an aircraft already equipped with it. That's not a requirement to get your instrument rating. After you get your rating, you are free to get one installed or go buy another airplane with it and there's no training requirement.

Also, ATC will not warn you if are about to hit the ground from blowing through the DH on autopilot. Their primary responsibility is aircraft separation, in terms of keeping aircraft from running into each other. DH is typically 200' above the runway. In those kinds of conditions, tower won't see the aircraft since visibility is so low. Approach control is the one looking primary at radar, but no one is going to notice anything out of the ordinary as you continue descending 500+ fpm past DH. In some situations, approach control and the tower look at the aircraft's position on the ILS very closely (simultaneous ILS approaches), but they are looking at lateral separation and course deviation, not descent past DH and by all accounts, your descent past DH on autopilot would look perfectly normal. It's only that last 100 feet or so where you do "the thing" to make it so touching the runway doesn't kill you. 500fpm impact is enough to kill you.

And relying on ATC to warn you of the ground when you are not landing works about as well as you think it would. In other words, landmark accidents where commercial airliners plowed into mountains led to more automation systems installed in the aircraft, namely TAWS.

There's no co-pilot when you are flying your bonanza, TBM or King Air.

There is a totally new level "hitting the ground avoidance" that has been designed into some aircraft now, already credited with saving some lives. A friend helped develop this system for the F-16. It's an automated system that will prevent the aircraft from being run into terrain/the ground. It will recover the aircraft if the pilot fails to.

I work in the business of air traffic surveillance. I was talking about commercial aircraft pilots, not GA pilots. However, of the several GA pilots I work with, all are not semi-unconscious automobile drivers. I think you'll agree that GA pilots take training courses to get their FAA certification, which content is far above what the general driving public gets when they take high school drivers ed; State DMV hand out driver's licenses like candy at Halloween. So I'd bet that if a GA pilot has an auto pilot system in the plane he is flying, he well understands the operation of it and would not let it fly the plane into the ground, but I'm sure there are dumbass GA pilots too (I've never met one though). Most people in general do not read their car's owner's manual. Most drivers have no idea what auto pilot means and think airplanes fly themselves, so they think Tesla's Autopilot can drive the car by itself. I've been in Model 3 with the Autopilot in use and it scares the shit out of me, but then again I take the act of driving quite seriously since my life depends on it. I ride motorcycles on the street (for well over 40 years), so I have a completely different perspective on the act of driving than the general automotive driving public.

I'll repeat myself. In my opinion, the people who want a car that can drive itself (and just not in theory) don't have enough interest in their own personal safety and don't take the act of driving seriously. If more people did take the act of driving seriously, we wouldn't be having a discussion on this topic, because it'd not be a topic. Pilots use use auto pilot as a tool to fly the aircraft not as a substitute to fly it.

Finally, the automation of air travel is completely different situation than automation of ground-based transportation. The FAA is taking great strides to make the surveillance of GA aircraft better integrated into the NAS (not really on topic here). But the point I've made in this thread over and over is that automating the ground-based transportation system of the US is a completely different situation than the nation's air traffic system. There really is no comparison between the two. The FAA is an independent third-party that keeps aircraft separated; there is no third-party that keeps automated ground vehicle traffic separated. The FAA keeps aircraft from crossing paths at the same exact location at the same exact time and there 3-axis of motion for an airplane to avoid another aircraft on top of the several redundant systems for collision avoidance past ATC. Ground-based vehicular traffic has to to cross paths; there is actually a name for for such locations... intersections. If cars were kept at the relative distances that the FAA keeps aircraft separated, automated ground traffic would be spaced at distances that road capacity would be even less than it already is. No one would tolerate it. The engineering solution to keep cars from being at the same location at the same time, is extremely difficult and is very very expensive.

[antzcrashing]

I am all for additional training required to operate a vehicle which has autonomous tech i.e. "a mind of its own". When I watch a video of the tesla FSD doing a 90 degrees right turn and spinning the steering wheel as it does it (driver confused as to if they should keep their hands ready, intervene or what, I am seriously concerned about this getting to the general public, and not zero concerned about it being in a pilot beta program. In general the regulations need to catch up to the tech. I would even advocate for submissions being required prior to them rolling out, similar to medical device industry.

[RM7]

Originally Posted by antzcrashing

So you are saying that the FAA doesnt require you to specifically learn about autopilot during your training? Ok maybe. But doesnt mean any training school worth salt wont train you on it.

To obtain a private pilot certificate in the United States, you must be at least 17 years old and complete 35 flight hours in an FAA part 141-approved school, at minimum

It's just as I said, only if your airplane used for the instrument check has autopilot. Mine didn't. Lots of people have instrument ratings that never got training on autopilot. After the checkride you can go buy an autopilot or aircraft with autopilot and there's no one to "check you out".

You don't have to go to an FAA 141 school, you can also obtain your certificate in Part 61 or 142. 142 is much more rare for the basic pilot certificates like private, instrument, etc.

[RM7]

Originally Posted by Efthreeoh

I work in the business of air traffic surveillance. I was talking about commercial aircraft pilots, not GA pilots. .

Well then you'd be talking about autonomous busses and taxis, stuff that operates "for hire" like commercial aircraft, not people "driving" their own autonomous cars.


You are right, regulation is lax. Regulation is the last thing that people seen to want or will vote for. It's a catch 22. People will be literally dying from their freedom, but they'd rather have that then regulation to keep them safe. It is what it is.

[antzcrashing]

Originally Posted by JamesNoBrakes

Well then you'd be talking about autonomous busses and taxis, stuff that operates "for hire" like commercial aircraft, not people "driving" their own autonomous cars.


You are right, regulation is lax. Regulation is the last thing that people seen to want or will vote for. It's a catch 22. People will be literally dying from their freedom, but they'd rather have that then regulation to keep them safe. It is what it is.

Unless autonomous cars (or planes) start killing people left and right. Look at telsa sentiment right after the first famous death when the truck crossed the road in front, sentiment was way down. Look at boeing sentiment after the 747 max had issues (software but not self driving). The approval ratings of these fluctuate and are not always rosy. But that said people dont always find more regulation as the best way to solve it.

[Efthreeoh]

Originally Posted by JamesNoBrakes

Well then you'd be talking about autonomous busses and taxis, stuff that operates "for hire" like commercial aircraft, not people "driving" their own autonomous cars.


You are right, regulation is lax. Regulation is the last thing that people seen to want or will vote for. It's a catch 22. People will be literally dying from their freedom, but they'd rather have that then regulation to keep them safe. It is what it is.

My problem with ground-based automation is society will install it at the lowest implementation cost possible, which means it will not work well and we all get to be Guinea pigs for big tech, while they try to get it worked out. Ground autonomous driving is never going to reach the safety levels of air traffic even throwing GA into the mix (even in Alaska too ). People thinking having machines drive cars around is going to improve safety are not aware of the complexities of it. Traffic is going to have to slow down, thin out, and personal travel will become less convenient.

[Efthreeoh]

Originally Posted by antzcrashing

Unless autonomous cars (or planes) start killing people left and right. Look at telsa sentiment right after the first famous death when the truck crossed the road in front, sentiment was way down. Look at boeing sentiment after the 747 max had issues (software but not self driving). The approval ratings of these fluctuate and are not always rosy. But that said people dont always find more regulation as the best way to solve it.

The Boeing 737 Max issue is overblown. And fixed. People are still flying.

[antzcrashing]

Originally Posted by Efthreeoh

The Boeing 737 Max issue is overblown. And fixed. People are still flying.

Every 747 max was grounded until Nov 18. Maybe it was overblown, but sentiment was enough for that to happen. Sentiment matters

[RM7]

Originally Posted by antzcrashing

Every 747 max was grounded until Nov 18. Maybe it was overblown, but sentiment was enough for that to happen. Sentiment matters

The other side of the coin is that Airbus has been doing far higher levels of automation for far longer, and done well. They too had a few accidents, although one of the worst was "showing off" the airplane for a demonstration, but they have a long history of doing it all successfully. Read about their flight control "laws" and how it works. It's not "flying" as most of us learned. Most of the aerodynamics and flight control movements we learned go right out the window in an Airbus.

Boeing is relatively new to this, although MCAS has been in other products, like the C-17.

Boeing's learning curve may have been steeper as far as flight control fly-by-wire automation, but the bigger problems appeared to be cost-cutting and disregarding safety, such as AOA disagreements and what it should do in those cases. Europe has mandated there be 3 AOA sensors, one of which may be "synthetic" based on groundspeed and other inputs, but to have the system just go into full stall-protection "dive at the ground"-mode due to ONE sensor without an AOA disagree was a very poor design.

Some of these concepts, like stick PUSHERS, have been used on these aircraft for decades and decades...

[Canuck335]

People would rather play on their phones vs actually driving.
Yeah people are going to die because of semi-autonomous cars.

But in the long run it'll be far fewer than the idiots driving around today.

The US has some of the highest road fatality rates in the world
https://en.wikipedia.org/wiki/List_o...ted_death_rate

So you can understand why US government agencies are fully behind automation
https://www.nhtsa.gov/technology-inn...ehicles-safety

It has nothing to with freedom of choice or driver training.
As usual, it's stupid people taking away that choice for the rest of us.

[Canuck335]

Originally Posted by JamesNoBrakes

Boeing's learning curve may have been steeper as far as flight control fly-by-wire automation, but the bigger problems appeared to be cost-cutting and disregarding safety, such as AOA disagreements and what it should do in those cases. Europe has mandated there be 3 AOA sensors, one of which may be "synthetic" based on groundspeed and other inputs, but to have the system just go into full stall-protection "dive at the ground"-mode due to ONE sensor without an AOA disagree was a very poor design.

Boeing made a very stupid, high school level coder mistake.
Then spent millions trying to deny it.
And lost billions when they couldn't.

[Alfisti]

Coder? No. They made a conscious decision to ignore engineers, any time a company does that it ends in tears. I am a boeing fan but that was unconscionable.

[antzcrashing]

Originally Posted by Alfisti

Coder? No. They made a conscious decision to ignore engineers, any time a company does that it ends in tears. I am a boeing fan but that was unconscionable.

Agree. An amateur software bug would not make it through the level of testing and scrutiny that these companies are subject to. They made a choice and hedged bets on training, but training cant fix it. They put a software solution on a mechanical design flaw, a huge pitfall.

[RM7]

Originally Posted by Canuck335

Boeing made a very stupid, high school level coder mistake.
Then spent millions trying to deny it.
And lost billions when they couldn't.

There's a lot more going on than coding mistakes. AOA disagree, redundant inputs, procedures to re-trim in the case of full nose-down that are against human factors/ergonomics. Other design issues. It's not just a computer bug.

[TSM330i]

The second we have driverless cars is the second that I don't give a crap as to what I'm driving in.

For the life of me, I can't understand why manufacturers are clamouring for this technology. If I'm not driving, that means I'm a passenger. What difference does it make what I'm sitting in then?

I will not own a car at that point, I'll just use a rented car or summon one to come to my house to pick me up and go where I need to go.

[dreamingat30fps]

Originally Posted by TSM330i

The second we have driverless cars is the second that I don't give a crap as to what I'm driving in.

For the life of me, I can't understand why manufacturers are clamouring for this technology. If I'm not driving, that means I'm a passenger. What difference does it make what I'm sitting in then?

I will not own a car at that point, I'll just use a rented car or summon one to come to my house to pick me up and go where I need to go.

Maybe you won't but I'm pretty sure many people will still want to own a car. The vast majority of people don't care how a car drives. My mother didn't even want to test drive her corolla before she bought it.

So if a large portion of the population don't care how a car drives, don't care for driving and just use their vehicles to get from point a to point b they are NOW in the same scenario you would be if cars were all self driving. They could right now just call an uber... aka a self driving car for all intent and purpose. However people still own cars.

Couple of reasons off the top of my head people would own cars even if they are self driving.

- Convenience. Not having to wait for car to pickup
- Comfort, maybe you can buy a nicer more comfortable car than the fleet offers
- Prestige. Plenty of people buy cars for the badge, same would apply.
- Costs. Maybe you "drive" a lot and the fleet costs would be higher than buying your own
- Rural areas where fleet cars are not easily available
- Businesses would still need vehicles

[Alfisti]

You won't see it in your lifetime don't worry. This makes space travel look like Grade 2 math.

[freakystyly]

Originally Posted by Alfisti

You won't see it in your lifetime don't worry. This makes space travel look like Grade 2 math.

Yeah, look at SpaceX and how far they've come. Tesla started off with a bang but it seems they haven't innovated anything in years.

[Efthreeoh]

Originally Posted by Alfisti

You won't see it in your lifetime don't worry. This makes space travel look like Grade 2 math.

Nail hit squarely on head...

[antzcrashing]

Originally Posted by Alfisti

You won't see it in your lifetime don't worry. This makes space travel look like Grade 2 math.

This issue is that the problem is over constrained. You basically cannot meet the fantastical ideas that people have for sleeping in their cars and to drive itself on worldwide roads in the presence of other non-autonomous cars. The problem will have to be simplified by infrastructure changes (dedicated lanes or underground roads), a mandate that all cars must be autonomous in certain regions, or people reducing their expectations.

[wdb]

Uber bails on autonomous tech. Could be a business decision, I expect they haven't had the greatest year.

https://www.cnet.com/roadshow/news/u...s-aurora-sale/

[antzcrashing]

Originally Posted by wdb

Uber bails on autonomous tech. Could be a business decision, I expect they haven't had the greatest year.

https://www.cnet.com/roadshow/news/u...s-aurora-sale/

It is a bombshell that they did. But I agree it is a business decision, they are probably going IPO soon and they don't want losing parts of the business