Self Driving Cars?

[ERD50]

What I was saying in my post above is that while responsible drivers may welcome a monitoring system that helps them to be safe, there are people who want to drive in an aggressive manner. See video below.

They will not buy such a system unless it is mandated on all cars, ...

By this logic, we should not have speed limits, or any rules-of-the-road whatsoever, because some people want to speed, or break the rules.

Those rules are mandated. Those people you describe can go rent some time on a track, or just suck it up. We don't need them killing innocent people.

-ERD50

 

[NW-Bound]

Some of the bad behaviors are already illegal. Yet, we have problems enforcing them. Even speed limit radars and red light cameras could be beat by legal loopholes. Yes, I want to see more red light cameras.

Some of the bad driving behaviors are difficult to define and enforce automatically by a computer mounted in the car. Therefore, you cannot make it forceful enough to disable the car.

You can help drivers who are conscientious and want to be helped. The jerks need more forceful methods. I am not sure how you do that, unless you have the computer drive the car. But that is against the driver being in control, which you like.

 

[ERD50]

....

You can help drivers who are conscientious and want to be helped. The jerks need more forceful methods. I am not sure how you do that, unless you have the computer drive the car. But that is against the driver being in control, which you like.

I only "like" the driver being in control until SDC proves the driver does not add any value. That is along ways away.

-ERD50

 

[Midpack]

I don't think it's a re-hash - you never 'hashed' it in the first place! You seem to reject this, but like above, I see no logic in your dismissal of this.

It's pure logic, where's the flaw?

Now I know why you and ____ keep repeating yourselves over and over and over, the driver monitoring hash went on ad nausea (I'm was guilty then, not any more) at page 27 almost three years ago...

Driver monitoring will help with accidents where inattention is primary. But the primary cause of many accidents has less to do with inattention. If you're roaring drunk, have poor visibility, are speeding like a maniac - inattention is far down the list. I suspect even many texters would ignore or turn off driver monitoring after the system arbitrary (not based on any hazard, just attention) warns them they're not paying attention AND they look up and see all is well. Texters are fully aware of the risks of inattention, and still choose to be distracted because they get away with it so often - until some don't.

But I know you'll remain unconvinced. No worries.

And your heady "logic" statement speaks volumes.

 

[ERD50]

Now I know why you and ____ keep repeating yourselves over and over and over, the driver monitoring hash went on ad nausea (I'm was guilty then, not any more) at page 27 almost three years ago. ...

It sure would help if you didn't say things like 'some posters', or blank out a name like you did above. What's with all the ambiguity? If someone said something you don't agree with, counter it directly. What's the problem?

"Page 27" doesn't help me - users can configure the number of posts per page, I have no page 27 for this thread (IIRC I'm set for 100 posts per page or whatever the max is, easier to do a 'find' w/o going back page by page). Got a post #?

NW-Bound certainly knows his stuff, he's no internet poser. He's related his work on an actual auto-pilot application! Disregard him at your own risk.

....

Driver monitoring will help with accidents where inattention is primary. But the primary cause of many accidents has less to do with inattention. If you're roaring drunk, have poor visibility, are speeding like a maniac - inattention is far down the list. ...

But I know you'll remain unconvinced. No worries. ...

I addressed that, but you just ignore it? If what I said has no merit, spell it out.

...
And your heady "logic" statement speaks volumes.

Then provide a counter to it. How can that simple logic not be true? What am I missing?

Maybe it would help to succinctly state what your thoughts are on SDC, for reference. Maybe it's got lost in a forest/trees way? Here's what I think you are saying (and please correct and add as needed):

What I hear you say is, people make mistakes when driving, or don't react fast/well enough. When SDC is better than humans, it will save lives.

OK, that's basic logic and it's fine as far as it goes. But IMO, it lacks any meaningful context. When, and what do we do until it is good enough to take over for a driver?

What has you so upset about the idea the we keep the driver engaged while we improve SDC?

I don't think you responded to Elon's predictions being totally blown out of the water. If Tesla is at baby steps four years after they said it would take two years to go cross country w/o a person on board, doesn't that give some indication that full SDC is very, very far out? Are you discounting Tesla's work, but counting on Waymo or others? If so, maybe we should start a thread specific to Waymo?

If I take a look into a blurry crystal ball, here's how I envision the next step in 'full' SDC (driver can drop all attention): There may be sections of highways approved for SDC mode. Maybe there would be an app that you need to be linked into, and that app would allow/deny full SDC depending on conditions. Clear weather, no unusual events, etc - you are cleared for SDC. Something happens and it alerts you, if you don't take the wheel, it tells the car to pull over.

Because obviously, full SDC is easier under certain defined conditions. So those are the first where it will be enabled. We don't have to wait for it to be good enough to deal with snow, road construction, cross traffic, etc. If it can perform well under good conditions, then let's start using it under those conditions.

Every tech I can think of has progressed along similar lines. It doesn't go from zero-to-hero overnight, it's a progression, and often is used differently during the progression.

-ERD50

 

[NW-Bound]

Now I know why you and ____ keep repeating yourselves over and over and over, the driver monitoring hash went on ad nausea (I'm was guilty then, not any more) at page 27 almost three years ago...

Driver monitoring will help with accidents where inattention is primary. But the primary cause of many accidents has less to do with inattention. If you're roaring drunk, have poor visibility, are speeding like a maniac - inattention is far down the list. I suspect even many texters would ignore or turn off driver monitoring after the system arbitrary (not based on any hazard, just attention) warns them they're not paying attention AND they look up and see all is well. Texters are fully aware of the risks of inattention, and still choose to be distracted because they get away with it so often - until some don't.

But I know you'll remain unconvinced. No worries.

And your heady "logic" statement speaks volumes.

Well, I don't know if I am the unnamed poster in the thread above, but it does not matter.

If someone really read what I wrote, then he would see that I agreed with Midpack that some bad behaviors cannot be helped with driver monitoring.

I thought that I made my thought very clear in many posts, so I will just quote the one immediately above Midpack's post.

Some of the bad driving behaviors are difficult to define and enforce automatically by a computer mounted in the car. Therefore, you cannot make it forceful enough to disable the car.

You can help drivers who are conscientious and want to be helped. The jerks need more forceful methods. I am not sure how you do that, unless you have the computer drive the car...

See!

How did I manage to get someone upset, when I agreed with him


Until we have the technology to remove the steering wheel, can some driver assistance functions help? I think so.

Automatic emergency braking, pedestrian detection, blindspot monitoring, etc... should all help. In theory, even when the system is not perfect, it does not eliminate accidents, but would reduce the accident rate I hope.

The only misgiving I have is that if the driver gets lazy and relies on the system and the system still has deficiencies, that could backfire. I guess time will tell how it works in real life.

We cannot eliminate the really bad actors unless we take away the steering wheel, but until we have full SDC should we not do anything?

Remember that some of the ADAS functions such as pedestrian detection, when it gets perfected, will find their way into the full-blown SDC. The development effort is not wasted. As the AAA test shows, the pedestrian detection in the Tesla 3 is currently so weak (worse than that of the Camry), the Tesla 3 autopilot would mow down pedestrians left and right if the driver does not intervene.

Again, these functions have to be a lot more reliable than they are now, before they can work in a truly autonomous SDC. But when used in ADAS, they can be quite imperfect and still save lives. Is that not obvious?

 

[NW-Bound]

If I take a look into a blurry crystal ball, here's how I envision the next step in 'full' SDC (driver can drop all attention): There may be sections of highways approved for SDC mode. Maybe there would be an app that you need to be linked into, and that app would allow/deny full SDC depending on conditions. Clear weather, no unusual events, etc - you are cleared for SDC. Something happens and it alerts you, if you don't take the wheel, it tells the car to pull over.

Because obviously, full SDC is easier under certain defined conditions. So those are the first where it will be enabled. We don't have to wait for it to be good enough to deal with snow, road construction, cross traffic, etc. If it can perform well under good conditions, then let's start using it under those conditions.

Every tech I can think of has progressed along similar lines. It doesn't go from zero-to-hero overnight, it's a progression, and often is used differently during the progression.

-ERD50

I have to remind myself often that SDC does not mean a completely autonomous car, and can have a different level of capabilities as defined by SAE.

What Midpack wants is the ultimate Level 5 (no steering wheel). You and I want that too. At least I made myself clear on that AGAIN and AGAIN.

But after that, we start to disagree. On what? You and I keep saying that it is not easy, and we will not have it that soon. What does Midpack think? That Elon has it already? No, it is far from a Level 5. Not what I have seen in numerous videos. Not what Tesla 3 owners say. Read what T-Al said about his recently purchased car.

Does me pointing that out make Midpack upset? I really don't understand.

And if we do not have it soon, is there something useful out of this we can have in the mean time? I think so. But the reduced capability has to be defined, so that we know the safe operating envelope of the current state-of-the-art. All products that we use have operating limits defined by the manufacturer. A tire has a maximum inflating pressure, and a load rating. A ladder has a load capacity. An SDC at less than the ultimate level 5 needs to have its capability defined.

Talk about any deficiency below Level 5 makes Midpack upset too. I really don't understand this. Level 5 would be wonderful, but we do not have it. Can we use the lower levels in SAE's definition? I think so. Why would anyone get upset over that? Should SAE abolish those lower levels because they are just useless?

It's like somebody insisting on taking a supersonic flight to Europe, but there's no supersonic commercial jet now. Do we still want to travel or not?

 

[Midpack]

Clueless...

 

[ERD50]

So here's a "small step" that I found interesting:

https://www.bizjournals.com/atlanta...e-push-with-gaussin.html?ana=yahoo&yptr=yahoo

United Parcel Service Inc. (NYSE: UPS) said it's testing the suitability and durability of Gaussin electric vehicles, known as "shifters," which have autonomous driving capability, to move semi-trailers and containers on the grounds of UPS’s advanced-technology London Hub.

Gaussin SA, a French engineering company, and UPS have been collaborating since 2018 to co-develop a competitive electric shifter vehicle. UPS says Gaussin electric vehicles have no tailpipe emissions and include "novel battery-swap technology," where discharged battery packs can immediately be replaced by fully charged ones.

So the autonomous gets a work-out in a more controlled situation, staying on the grounds of UPS. That should help in collecting data. So far, the videos I've seen of Teslas in "summon mode" in parking lots look pretty bad. Still a lot to learn apparently.

These are electric, and the battery swap is a good idea I'd say. That way, you aren't under constraints for a fast charge, which takes a huge amount of power. They can charge a spare battery slower - takes roughly the same energy, but lower instantaneous power.

-ERD50