Self Driving Cars?

I am sure that car developers are careful, but am afraid that in the rush to "beat the competition", they are going to cut corners. And then, the public, being so used to the attitude that says "Ah, just reboot it and it will be fine", will not realize that a deficiency in software or hardware will reboot them to the next life (if they believe in reincarnation that is). :)

Yes. Beat the competition, do it cheaper, and so on. This is a concern. There is a lot of pressure to push out things on a prayer. See it happening all too much in many industries.
 
I agree too, except when he said to build-in security later!! :facepalm: Dontchaknow that's why we're total screwed, security-wise, with just about everything nowadays?

Question... Does anyone know about a transponder proposal for inter-car communication? It would seem to me that all cars, not just those with automation, could send out alerts when the car has to swerve or hit the brakes (or maybe even long horn blasts, indicating rage...see other thread). Any automated cars would "hear" the alerts that came from the path ahead and signal the alarm to either the real or automated drivers in the area.

It is being worked on. But nothing is set in stone, and there are differing ideas. This is one of the many things that needs to be resolved.

Vehicular Communications Systems

Some of the conflict over spectrum is troublesome. I'm not sure why some think that autonomous cars would make the need unnecessary. What?

Internet service providers (who want to use the spectrum) claim that autonomous cars will render V2V communication unnecessary. The US automotive industry has said that it is willing to share the spectrum if V2V service is not slowed or disrupted; and the FCC plans to test several sharing schemes.
 
I am sure that car developers are careful, but am afraid that in the rush to "beat the competition", they are going to cut corners. And then, the public, being so used to the attitude that says "Ah, just reboot it and it will be fine", will not realize that a deficiency in software or hardware will reboot them to the next life (if they believe in reincarnation that is). :)
It's possible. However, the potential liability costs and long term damage to the brand if they overlook issues that put riders in significant danger should also discourage cutting corners. They have strong incentives to get it right. Always trade offs.
 
Some of the conflict over spectrum is troublesome. I'm not sure why some think that autonomous cars would make the need unnecessary. What?
I do not follow all this, but think that true autonomous cars without V2V communications still make the road safer than it is now. The V2V provides the enhancements, such as making traffic flow more effective.

Cars travelling bumper-to-bumper and merging like zippers at 70 mph on congested freeways. Cars weaving at city street intersections without slowing down. I think people envision all this good stuff that requires V2V for coordination.

But man oh man, one hiccup and you have a chain reaction. Car tires do blow out, hardware does fail. I dunno.
 
It's possible. However, the potential liability costs and long term damage to the brand if they overlook issues that put riders in significant danger should also discourage cutting corners. They have strong incentives to get it right. Always trade offs.
In all levels of automation below Level 4, the driver is ultimately responsible.

They cannot tell you if the system can see a roadside obstacle or not, or can swerve to make room for a bicyclist, etc... Yet, everything else much less critical on a car has a warranty spelled out: battery good for 60 months, tires good for 75K miles, etc...

And as there is no clearly defined level of performance of the "autopilot", what it can see and avoid, the car makers always get a "get out of jail free" card. It is only at Level 4 or 5 that you can hold the maker responsible.

Yes, in the long run the truth will out, but this makes me not at all willing to be the first guinea pig.
 
Will it be that different from buses and taxis? Millions of people use them everyday.

Even in cities with great public transportation, where it's usually the most efficient way to get around, people of means will buy cars, just so they don't deal with crowds and potential hygiene and crime.

Now if self-driving cars are priced like what people now pay for a town car with a driver in a city like NY, then there will be few people who can afford them.

But self-driving cars are being developed by car makers which produce cars for the masses, not Bentley and Rolls Royce.
 
It's possible. However, the potential liability costs and long term damage to the brand if they overlook issues that put riders in significant danger should also discourage cutting corners. They have strong incentives to get it right. Always trade offs.

Which is why IMHO the aviation solution will be adopted the driver will have the ultimate responsibility just like the pilot. Yes the automation may make a mistake, but it will be the drivers fault, just like it is the pilots fault (In SF crash the pilots selected the wrong mode for the automation, and did not monitor close enough).
 
Which is why IMHO the aviation solution will be adopted the driver will have the ultimate responsibility just like the pilot. Yes the automation may make a mistake, but it will be the drivers fault, just like it is the pilots fault (In SF crash the pilots selected the wrong mode for the automation, and did not monitor close enough).
:nonono: Not at all true. :nonono:

Autopilot and autothrottle modes are always announced on the display panels, in case the pilots forget. In that accident, the pilots selected an autopilot mode that resulted in the autothrottle getting turned off by design. I am sure that was annunciated in case the pilots forgot what was supposed to happen. But they were too busy to look at the displays probably. By the way, displays are always at least dual, in case one fails.

In all the systems I know, when the autopilot is armed for an automatic landing, if it announces "Autoland", then it means every subsystem is a go. That means all sensors critical for autoland are working and in agreement, such as multiple navigation receivers working and in agreement (glide slope and localizer receivers), at least dual radar altimeters, etc... Multiple (2 or 3) computers must also be in agreement, etc...

These are hard requirements! If the system fails, it will abort the Autoland, annunciate it, and the pilots have to land by hand. If a hardware failure occurs prior to the Autoland mode being selected by the pilots, it will not engage the mode and tell them so. It's the FAA rule. The pilots are there in case one of the sensors or navigation receivers fail and the automatic pilot does not have sufficient redundancy for a safe landing.

The reason for the dual and triple sensors is this. Any single thing can fail, but it is nearly impossible for two or three to fail simultaneously, and putting out exactly the same wrong information. So, by comparing them to verify that they agree, you can be sure that they all work.

Once one sensor of a pair has failed, the remaining one cannot be trusted 100%. Flight modes that are not critical can still use it, but Autoland is a nono. So, really critical sensors are triplex in order to still have a working pair if one of the 3 fails.

PS. In the early days, some of the indicators are just light bulbs. Now, light bulbs can burn out. So, the indicators have two light bulbs in it, and they are lighted by two different paths (a single wire can get cut or breaks for example).

When you push a button, there are actually two switches behind that, in case one fails to make contact. Again, there are separate wires going to each switch. It's serious stuff.

PPS. It appears that pilot's training was inadequate. The airlines are responsible for keeping their pilots up-to-date, and to send them for recertification or to refresh them on how to fly a particular aircraft.
 
Last edited:
Not everyone will be a hazard. For example there are plenty of drivers that can stop faster than the best current ABS brake technology.

I expect the same thing will happen with self-driving cars...it will take a very long time for a self-driving car to meet or exceed the capability of a highly skilled, professional driver.
Actually the self driving systems will prep the car for the accident, and at a minimum automatically brake (because lower speed always means less impact even in a head on crash) I do assume that self driving cars will be equipped with all the air bags a modern car has. Note that with the average age of a car in the US fleet now being 11.4 years It will be 15+ years from the time the cost of a level 5 car gets comparable to mid level other cars before any talk of regulation could occur. I suspect the first thing might well be autonomous car only lanes on freeways (which are the most controlled driving environment anyway) sort of a step up from express lanes. The less like a freeway the harder the self driving problem is as more things change. One question, assume a big dog runs in front of the self driving car, what happens.
 
Yes, we are all conjecturing at this point. Predicting where technology will succeed and where it fails is hard. I already brought up the SST Concorde and the Space Shuttle.

But while we are talking about the possibility in the future (is this not the case as we have not seen a Level 5 car yet :) ), in the case the technology has not driven down the cost of Level 5 car down for mass consumption by the time you become frail and invalid, would you buy it at $150K? At what price would you "go for it"? Remember that you do not have decades to wait.

Actually if you live in areas served by Uber you can get uber delivery, and many grocery stores will pick you order waiting for pickup. This also applies to takeout. Again depends on where you live I live to far out for Pizza delivery so I do have to pick it up.
 
:nonono: Not at all true. :nonono:

Autopilot and autothrottle modes are always announced on the display panels, in case the pilots forget. In that accident, the pilots selected an autopilot mode that resulted in the autothottle getting turned off. I am sure that was annunciated in case the pilots forgot what was supposed to happen. But they were too busy to look at the displays probably. By the way, displays are always at least dual, in case one fails.

In all the systems I know, when the autopilot is armed for an automatic landing, if it announces "Autoland", then it means every system is a go. That means all sensors critical for autoland are working and in agreement, such as multiple navigation receivers are working and in agreement (glide slope and localizer receivers), at least dual radar altimeters, etc... Multiple computers must also be in agreement, etc...

These are hard requirements! If the system fails, it will abort the Autoland, annunciate it, and the pilots have to land by hand. It's the FAA rule.

PS. In the early days, some of the indicators are just light bulbs. Now, light bulbs can burn out. So, the indicators have two light bulbs in it, and they are lighted by two different paths (a single wire can get cut for example).

When you push a button, there are actually two switches in there, in case one fails to make contact. Again, there are separate wires going to each switch. It's serious stuff.
As indeed some of the newer features have buttons that show their status in cars, such as lights showing lane departure warning and parking assist. Autonomous cruze control also provides lights (indeed regular cruse control does also)
 
Well stated. Convenience/where one lives will be a critical variable. But so is cost. We can discuss all day and never reach a solution without having a common assumption on (relative) cost, and there's no solid basis on cost. If level 5 cars come to pass in sooner and at a lower premium to level 0-2 cars, more people will own them directly. But if level 5 cars take decades and start at a considerable premium, we may all have been forced to adapt to sharing, and won't need to own cars. The 'rural person who needs to observe their fields' will have already bought an ATV, golf cart, or a drone for that purpose. Or there might be real time Google Earth by then. We'll all weigh the cost of sharing vs owning, our $ resources, and our "convenience" needs/desires - and make our choice.

It's an aside to the above, but some posts are inadvertently further confusing the issue because we don't have common assumptions of self-driving - I was guilty of it not long ago. Some people are referring to level 3-4 cars, where others mean mature level 5 cars (the frame of reference I'm trying to work from).

Since one factor of cost is the value of a persons time, and that varies greatly that is something that can't be agreed on (The issue here starts with wait times for a summoned car to appear)
 
As indeed some of the newer features have buttons that show their status in cars, such as lights showing lane departure warning and parking assist. Autonomous cruze control also provides lights (indeed regular cruse control does also)

The difference between commercial aviation and automobile is how reliable the indications have to be.

A car uses a vision camera to look at the lane markings. Does it have dual cameras to make sure they see the same thing? Is each camera interfaced to a different computer to make sure the images are interpreted the same?

Surely all cars have cruise control, but do they have dual speedometers? Wired up to dual computers? Having dual actuators?

I am not saying a car has to be as fail-safe as a commercial jetliner, but these are the differences. And that's how a jetliner costs hundreds of million.
 
Last edited:
Note that riding a horse is legal by state law in all states except on roads labeled motorized vehicles only, i.e. freeways. Even in NYC you still have the folks that provide horse drawn rides around Manhattan.
 
I've got a Tesla Model X on order, with all the automated driving features, most of which are yet to come. That includes "full self driving". This is a car we'll be driving on cross-country trips. the more driving it can do the easier that will be.
<snip>
I do expect this to be my last car purchase. I assume autonomous car sharing or taxis will be more economical for occasional use in the future. So I'll be renting or subscribing or something different then.

I have been driving a Tesla Model X (SUV/CUV) for just over 1 year and just turned over 27K miles today. I probably have driven 90%-95% of those 27K on Auto-Steer/Pilot.

Yes, I've taken multiple road trips to get that many miles and this is one place the car shines. It stays steady centered in the lane and with a 2.5 second delay behind the car in front of me (if there is one). It makes for a MUCH more relaxing drive. You still need your hands on the wheel but you can look around more and have better conversations.

I also use it a lot around Chicagoland and even on 35-50 MPH roads. Why? It is looks at TWO cars ahead with the latest changes of using radar primarily and the camera secondary. So even when a van is in front of me it detects the little car in front of it (van) slowing down or braking even if the van brake lights have not come on yet. It is also very steady in the center of the lane and keeps a safe distance (again 2.5 seconds).

FYI, I use the maximum following distance unit of 7 which roughly translates to 2.5 seconds (so more distance at higher speeds and less distance at lower speed).

I drove 100+ miles today to a sporting event where 70% was on the highway and 30% was on 35-55 MPH roads. The Tesla Model X drove 95% of this 100+ miles. Relaxing drive.

ASIDE and FYI ---- EVERY TIME you engage autopilot a message pops up to remind you to KEEP YOUR HANDS on the wheel. If you don't and ignore it's warnings 3 times THEN you have to pull over and put the car in park before you can then drive and re-engage autopilot.

Below image is agreement I had to click on to enable AutoPilot -- done at walkthrough.
fz7Fiy5.png
 
Last edited:
The difference between commercial aviation and automobile is how reliable the indications have to be. ...

In some ways, isn't car software far more complicated?

A plane rarely has traffic nearby. No chance of a kid on a bike suddenly popping out between two parked cars. Nobody pulling out of a parking space, opening a door into traffic, running a red light. You aren't travelling with other planes so close that they could go from being just fine one second to crashing into you in the next second. We see chain reaction pile ups with cars & trucks - you aren't going to see that with planes.

I suppose some of those things do happen with a plane, but they can happen to me dozens of times on even a short trip into town. Seems far more challenging.

-ERD50
 
In some ways, isn't car software far more complicated?

A plane rarely has traffic nearby. No chance of a kid on a bike suddenly popping out between two parked cars. Nobody pulling out of a parking space, opening a door into traffic, running a red light. You aren't travelling with other planes so close that they could go from being just fine one second to crashing into you in the next second...

Definitely!

In an earlier thread, I stated that a car driving software requires far more sophistication and faster CPU than flying an aircraft. Hence, I was both impressed at what has been done, and alarmed at the same time at the cavalier attitude I have seen. We took a lot more care, and our product was used by professional pilots, not dummies.
 
Last edited:
... ASIDE and FYI ---- EVERY TIME you engage autopilot a message pops up to remind you to KEEP YOUR HANDS on the wheel. If you don't and ignore it's warnings 3 times THEN you have to pull over and put the car in park before you can then drive and re-engage autopilot...
How long can you keep your hands off the wheel until the computer warns you?

On the Web, I learned that it was a recent software update. Earlier, a guy on the Web made an experiment to see how long he could keep his hands off, and found that it varied. It could be up to 10 minutes one time, and a lot shorter other times. Seems like the sensing mechanism is not a fool-proof thing.
 
OK. I might have found the answer. Some Tesla owners try to reverse engineer that sensing mechanism. They are convinced that there is no sensor inside the steering wheel.

Rather, the servo driving the wheel can sense the resisting torque (torque sensor on the shaft?) when the driver's hand is on the wheel. But for it to sense any resisting torque, it has to move the steering wheel, and it only does that on a curve road. So, on a straight road you can fool it.

This guy wrote the following. Note that I highlighted the bolded sentences.
There are no sensors in the steering wheel itself, I believe.

I hold the wheel lightly with the fingers of both hands, to
be able to feel the normal tiny corrections in steering that
the AS is making. Then, I can feel any peculiar steering
action, even before the car has a chance to move over
very far.

Recently that instant feedback seemed to be instrumental
in "saving our bacon" (preventing a serious accident) when
our MS appeared to suddenly fall in love with some yellow crash
barrels that we were going to pass. It tried to swerve toward
them
, instead of just driving smoothly past in the exit-only lane.

Unknown why, because the two exit only lanes were well marked,
and we had been in the leftmost of these two long lanes for
perhaps 1000 feet, or so. Dry, clear weather, no glare, etc.

If my hands had not been on the wheel, I would have been too
late, or so it seemed
.​
 
Last edited:
How long can you keep your hands off the wheel until the computer warns you?

On the Web, I learned that it was a recent software update. Earlier, a guy on the Web made an experiment to see how long he could keep his hands off, and found that it varied. It could be up to 10 minutes one time, and a lot shorter other times. Seems like the sensing mechanism is not a fool-proof thing.
There are multiple levels of warnings. Message, then flashing white light around the edge of the drivers display, then noises that get progressively "madder" :) . It is that 3rd level that you can't ignore more than 3 times before you have to park.

It is not a simple time thing because it looks at the traffic around you and what the roads are like (curves). It is fairly aggressive and I get the 1st or 2nd level regularly. Part of that reason is EVEN tho my hands ARE on the wheel it only uses the steering wheel rotational torque as an indication your hands are on the wheel. So if you grip lightly (resistance) or are on VERY straight roads you get warnings.

I think it will come on less than 1 minute but varies based on road and traffic conditions ... and perhaps speed you are traveling.
 
OK. I might have found the answer. Some Tesla owners try to reverse engineer that sensing mechanism. They are convinced that there is no sensor inside the steering wheel.

Rather, the servo driving the wheel can sense the resisting torque when the driver's hand is on the wheel. But for it to sense any resisting torque, it has to move the steering wheel, and it only does that on a curve road. So, on a straight road you can fool it.
You replied while I was typing ... BUT keep in mind if you hold on fairly tight then the RESISTANCE of the wheel turning lets it know you have your hands on the wheel.

For example, when I'm doing long highway travel, I can rest my arms on the armrest or legs but still hold the wheel fairly tight and REIST the turning "enough".
 
Will it be that different from buses and taxis? Millions of people use them everyday.
A significant difference is that buses and taxis are semi-public accommodations with human attendants. There are some social checks on the bad behavior that might occur and there is someone to notice and clean it up, or at least request a cleanup, when it occurs. Compare that to the debacle of self-cleaning restrooms installed in some cities. They quickly became unusable. No one was watching to see what kinds of drug deals or debauchery occurred in the privacy of an automatic restroom. They were quickly destroyed.
 
You replied while I was typing ... BUT keep in mind if you hold on fairly tight then the RESISTANCE of the wheel turning lets it know you have your hands on the wheel.

For example, when I'm doing long highway travel, I can rest my arms on the armrest or legs but still hold the wheel fairly tight and REIST the turning "enough".

There are already guys trying to defeat this warning. One said he could hang a weight on the steering wheel and it worked.

That, in a thread where an earlier guy in the same thread said having his hands on the wheel saved him when his Tesla suddenly veered towards yellow construction barrels. It has an affinity for the yellow color, hmm... See my edited post above.

It is hard to make anything fool-proof because fools are ingenious.
 
Last edited:
Back
Top Bottom