Self Driving Cars?

[Midpack]

True, but I think we should keep focused on the goal - reducing accidents/injury/death. If the SDC can detect things we would miss, that's a positive, and could help bring down the averages. We shouldn't only think in terms of replicating what humans do. That kind of thinking is too narrow, like designing robots with only two hands, when three or more can be more effective.

Along those lines, I feel that while the pursuit of Level 4-5 is interesting and will have benefits along the way, I'm a more pragmatic thinker (the good in the near term over the perfect future). I fall back on the 80/20 Pareto Principle, we should be looking at the biggest problems (some of that has been reported in this thread), and looking at what current and near term technology can do to help in those areas. What SDC team isn't prioritizing their efforts? So they set out to solve the original 80. Unless that effectively eliminates accidents/injuries/fatalities, it just sets up a new 80/20 from the original 20. When innovation can proceed cost effectively, development continues.

Just to illustrate with numbers, if a 30% problem is difficult and expensive to detect/respond so is only 5% effective, but the 10% problem is relatively easy and 90% effective, there is more gain in that low hanging fruit - 9% solved easily, versus 1.5% solved at high cost.

Thought provoking, interim thinking?

Reducing or eliminating accidents/injuries/fatalities is the goal. What SDC team (Waymo et al) isn't working toward that?

The assumptions that would have to be underlying your POV?:

• You believe accidents/injuries/fatalities can be almost eliminated (90% human error) without going all the way to level 4-5. That would be less costly, less disruptive, and might happen sooner. But it appears many in the auto-tech industry, who know way more than we do, believe we're not going to get there without reaching for level 4-5 - and they're investing big $ in that pursuit.
• And you believe some lesser level car is on a wholely different development path. I guess that's possible, but it seems counter-intuitive. I am sure the SDC teams know better than we do what tech is complementary and what may not be.

Those (few?) in the auto-tech industry who aren't committed to level 4-5, may test your theory. Time will tell.

I assume the SDC groups are far more knowledgeable than we are. A couple prolific posters here seem to assume they aren't...

 

[ERD50]

Thought provoking, interim thinking?

Reducing or eliminating accidents/injuries/fatalities is the goal. What SDC team (Waymo et al) isn't working toward that?

To summarize the assumptions underlying your POV:

• You believe accidents/injuries/fatalities can be almost eliminated (90% human error) without going all the way to level 4-5. That would be less costly, less disruptive, and might happen sooner. But it appears many in the industry, who know way more than we do, believe we're not going to get there without reaching for level 4-5 - and they're investing big $ in that pursuit.
• And you believe some lesser level car is on a wholely different development path. I guess that's possible, but it seems counter-intuitive. I am sure the SDC teams know better than we do what tech is complementary and what may not be.

Those (few?) in the industry who aren't committed to level 4-5, may test your theory. Time will tell.

I assume the SDC groups are far more knowledgeable than we are. A couple prolific posters here seem to assume they aren't...

That's not a very good summary of what I've been saying, or at least thinking (any communication breakdown might have occurred on my end), so let me try to clarify...

For #1 in your list, I never said almost all human error could be eliminated with driver assistance/monitoring/engagement. I don't know the numbers, I've only been saying that I think monitoring the human involvement along with some driver assistance is better than driver assistance alone. That hardly seems even questionable to me. X+Y must be > X (for positive values!).

I also think, and I believe said along the way, that the investments might be more due to the 'sexy' nature of self driving, rather than what is best for the end user. We see that all the time in marketing, don't we?

For #2 in your list, I did not say that at all. I said they can work together - keep the driver involved and engaged, and at some point, if we almost never need the driver to intervene, we are 'there'. A see them as complementary, and converging along the way (and maybe not even needed in the future).

OK, the SDC teams are working towards better safety. It is my reasoned opinion that we would be better served with systems that focused more on human involvement/engagement/monitoring to work with the driver awareness and SDC features. They can supplement each other better than either one can work alone, for a very long time to come. I just feel that the current focus on relieving the driver isn't the right path.

Again, the SDC teams are more knowledgeable, but if their goal is 'sexy' marketing, and maybe some specific long term profit goal, those goals might not really be serving the public in the best way possible, especially in the near term.

I can go back and find my exact quotes if you wish, I'm going by memory, but I think I covered my earlier points accurately.

-ERD50

 

[Midpack]

That's not a very good summary of what I've been saying, or at least thinking (any communication breakdown might have occurred on my end), so let me try to clarify...

For #1 in your list, I never said almost all human error could be eliminated with driver assistance/monitoring/engagement. I don't know the numbers, I've only been saying that I think monitoring the human involvement along with some driver assistance is better than driver assistance alone. That hardly seems even questionable to me. X+Y must be > X (for positive values!).

I also think, and I believe said along the way, that the investments might be more due to the 'sexy' nature of self driving, rather than what is best for the end user. We see that all the time in marketing, don't we?

For #2 in your list, I did not say that at all. I said they can work together - keep the driver involved and engaged, and at some point, if we almost never need the driver to intervene, we are 'there'. A see them as complementary, and converging along the way (and maybe not even needed in the future).

OK, the SDC teams are working towards better safety. It is my reasoned opinion that we would be better served with systems that focused more on human involvement/engagement/monitoring to work with the driver awareness and SDC features. They can supplement each other better than either one can work alone, for a very long time to come. I just feel that the current focus on relieving the driver isn't the right path.

Again, the SDC teams are more knowledgeable, but if their goal is 'sexy' marketing, and maybe some specific long term profit goal, those goals might not really be serving the public in the best way possible, especially in the near term.

I can go back and find my exact quotes if you wish, I'm going by memory, but I think I covered my earlier points accurately.

-ERD50

No need, your clarification isn't objectively at odds with my summary.

Again, I am sure the SDC teams have thought about your ideas far more thoroughly than you have. Driving assistance is largely a subset of full SDC. They may get results without pursuing full SDC, though it's clear Waymo and others think it'll take level 4-5 to get there. As many others have pointed out, the more tasks the car takes on, the less attentive drivers may become. That would suggest accident stats for level 3 and below could get worse if development pauses there - but I don't pretend to know for sure. The SDC teams undoubtedly know better.

If your asserting cars can be more assistance capable and enhance driver awareness, that seems counterintuitive too. More lights, buzzers, vibrations, audibles and alarms won't work. People won't buy that, and if mandated there will probably be a cottage industry to defeat them.

Again, I know the SDC teams have a much, much better handle on this than we do. A couple prolific posters believe otherwise, and want us to believe their views carry more weight than others.

And again, this is an interesting thread and I've learned a lot, even from the 1-2 prima donnas! Thanks.

 

[NW-Bound]

Midpack, it is so sad to see you still not understanding what I have been trying to say.

Misconceptions:

The first models of fully autonomous cars will be targeted to the consumer and will be available for purchase.

Public demonstrations of self-driving cars provide an indication of their capabilities

I could not agree more with the two above misconceptions that you quoted.

About the 2nd in the above list, the writer points out the complexity of the problem if anyone cares to read what he wrote. So, I shall include the summary here.

To summarize: It is impossible to judge the maturity of a self-driving car by observing public demonstrations. Difficult situations don’t occur that frequently and therefore these demonstrations can only confirm that a prototype has reached quite a basic level of capability. The enormous difference in maturity between, for example, Google’s prototypes – the current leader in this technology with nearly two million kilometers of testing in autonomous mode and more than 10.000km of testing being added every week! – and the prototypes of all other developers of autonomous car technology can not be appreciated by observing public demonstrations.

I have been trying to say the same. Yet, you agree with him and not me. Why?

I shall try to repeat one more time (my wife calls me impatient, but I disagree). Car makers know about these problems very well. I point out the problems they are trying to solve, so that people appreciate what they are doing. I did not say anything is impossible, though some might be. Those hard problems are the difference between Level 4 and Level 5. Waymo is exploring to see what they can do.

I like Waymo a lot, because they try to educate the public about the problems they are trying to solve (some bragging is involved on their part too, but that is to be expected when one is so good compared to his competitors). Other makers, though they know of the problems, are not as visible in the public about this. I think it is because they are not as advanced, hence keep mum.

Therefore, I like to watch Waymo more than the other guys to see what new things that Waymo is doing. All other guys are playing catch up. I highly encourage people to watch Waymo public talks, of which we both have provided links to.

Communications can be so tough. One says 2+2=4. The other one says, no, 2x2 = 4, and disagrees with the 1st.

 

[ERD50]

No need, your clarification isn't objectively at odds with my summary. ...

Well, let's try to focus down to a few points to see if we can sort through it.

Again, I am sure the SDC teams have thought about your ideas far more thoroughly than you have.

I'm sure too, and I addressed that. Their goals may be different. It's the FA argument all over again - that FA might actually know more about finances than you or me, but their primary goal is not to increase our NW, their goal is to keep us as a customer and increase their NW.

If your asserting cars can be more assistance capable and enhance driver awareness, that seems counterintuitive too.

I don't see it as counter-intuitive at all. How so (I'll address the social issue later)?

I've described it several times over. SDC start out with limited capabilities. They will require driver awareness, pretty much continually (just like the warnings on the Tesla tell the driver, though they somewhat negate that with the term "Auto Pilot"). But we have seen that some drivers are lulled into a sense of security, and let their attention drift. If the SDC requires driver intervention (and it might not be smart enough to even know it needs help), and if that overly-complacent driver is looking at his phone instead of the road, that's a problem. How can it be counter-intuitive that an SDC system that is also monitoring the driver's eyes and head movement (and/or other signs) would be better than SDC alone? Again X+ Y > X for positive values.


OK, on social engineering:

More lights, buzzers, vibrations, audibles and alarms won't work. People won't buy that, and if mandated there will probably be a cottage industry to defeat them.

For this discussion, I was focusing on what technology could do. Whether it is accepted or not is another story, one that I don't have an answer to (and I don't think anyone does). But I'll think about it now -

First, you are wrong (strong un-wiggly word, but 100% the case) that people won't buy it. My frugal, LBYM self just paid $$$$ extra on my new car for lights, buzzers, audibles and alarms to warn me of cars in my blind spot, my approaching a car in front too fast, lane departure, and objects behind me when I'm in reverse. I found lots of articles online helping people find cars with these features, there's clearly a market for these products, and the car makers are providing products for that market. So you are wrong.

My motivation was if any of these features help me to avoid, or reduce the severity of one accident in the time I own the car, they were worth it. And like insurance, you sometimes appreciate having it, and never using it.

You don't pay to defeat a product you voluntarily purchased.

Another motivation - maybe insurance companies will offer discounts based on these features? As costs come down, and capabilities improve, that could be very positive.

I won't try to go too far to hypothesize on what would happen if mandated. I kinda think that things people object to won't get mandated. Back up cameras are being mandated, not everyone would want/pay for them, but I don't think too many people object to them, and fewer still (no one that I've ever heard of) would defeat them. I have not done a deep dive into the numbers, but backing up is dangerous, lives are lost, so it probably is a sensible mandate. Sometimes govt does things right - though of course you will never get 100% agreement on anything.

-ERD50

 

[Midpack]

No thanks. We disagree on your semi-coherent assessments above, SDC teams like Waymo analogous to an ordinary FA - you must be kidding?

You want to question the bonafide industry experts who clearly know way, way more that you do. I question the two self-appointed experts here who've repeatly confessed they don't know - yet the bulk of their narratives belies that confession. Carry on...

 

[ERD50]

No thanks. We disagree on your semi-coherent assessments above, ...

Well, at least I earned a "semi-coherent"!

SDC teams like Waymo analogous to an ordinary FA - you must be kidding?

You must be unfamiliar with term "analogy".

You want to question the bonafide industry experts who clearly know way, way more that you do.

I could not have been any clearer that I am not questioning their knowledge and expertise, I am questioning whether their goals are in alignment with what I see as positive interim goals. It doesn't make them 'stupid', it just means they have their sites set differently, and I made a clear case as to why what I described could be of greater benefit in the interim. X + Y, remember?

Your inability to understand that does indicate what I've felt for a while, but hoped against hope - there is no point in continuing this 'discussion'. And if you think referring to those with a different viewpoint from you as "prima donnas' helps your case, well, you just go on believing that.

-ERD50

 

[NW-Bound]

Independent of what people here say, myself included, someone who is really interested in this technology, not just someone waiting to buy a self-driving car in the next few years, will do well to watch Waymo public talks. The article that Midpack quoted is also good.

I do not claim to be an "expert" in this field, but I have learned that demos are like bikinis.

“Statistics are like bikinis. What they reveal is suggestive, but what they conceal is vital.” -- Aaron Levenstein

Yes, a maker shows what its car can do in a demo in a nice clean scenario. He is not going to show what his car cannot do. It's common sense.

So, a guy like Waymo who has got more to show than his competitor is more advanced. It is quite impressive. And when he says he is shooting for level 4, you have to wonder what is missing for level 5. Surely, he knows more than you, and he is not telling you what he still needs.

So, you wonder what his remaining problems are. You do not have to be an expert. Just common sense, really.

 

[HFWR]

I don't think many average Josephines are clamoring for self-driving cars. Seems to me the early adopters would be "fleet" oriented, such as cabs, buses, mail/package delivery, etc. Only that kind of volume would drive the manufacturing costs down, and trustworthiness up, to the point where regular folks could afford or would want one. Given the technological challenges, legal challenges, infrastructure challenges, et al, I think thirty years is overly optimistic for widespread adoption.

 

[Midpack]

I could not have been any clearer that I am not questioning their knowledge and expertise, I am questioning whether their goals are in alignment with what I see as positive interim goals.

Exhibit A.

Your inability to understand that does indicate what I've felt for a while, but hoped against hope - there is no point in continuing this 'discussion'.

Hello? I concluded that hours ago, you didn't notice my cursory replies didn't engage your POV?

Independent of what people here say, myself included, someone who is really interested in this technology.

Me too. I'm amazed at how far SDC's have come already.

So, you wonder what his remaining problems are. You do not have to be an expert. Just common sense, really.

We can wonder about the timeline, societal impact/changes, basic tech and the like. But IMO it's beyond presumptuous to think we know the tech details better, or to presume the SDC teams don't know development path/goals & objectives far better than we do. I seriously doubt any of what's appeared here has escaped them, more likely they've solved (hundreds of) thousands of problems that will never occur to us - and not just purely technical. If we disagree, all good, only time will tell. Many of your posts have been thought provoking, thanks.

 

[explanade]

If it's 30 years out, then it won't matter to many of us.

I don't think these companies are expecting that long a period for turnover.

 

[NW-Bound]

We can wonder about the timeline, societal impact/changes, basic tech and the like. But IMO it's beyond presumptuous to think we know the tech details better, or to presume the SDC teams don't know development path/goals & objectives far better than we do...

No, it's not presumptuous. I don't know why you keep bringing that up.

Exactly because they know more than I do that I wonder what their secrets are. When I cited these difficulties, it is the difficulties that I, as a layman in this technology, see it. If anything, it is so that people know how tough it is.

The more you think about something, the more you appreciate what someone has done.

 

[JoeWras]

Independent of what people here say, myself included, someone who is really interested in this technology, not just someone waiting to buy a self-driving car in the next few years, will do well to watch Waymo public talks. The article that Midpack quoted is also good.

Agree, the article was good. It does tip to optimism, but also has a lot of real world points. I think the part about first rolling out in select well mapped, controlled areas in commercial adventures is right on. I'm starting to think Über has it right! I thought they were crazy, but maybe not. Roll out SDCs among the regular traffic in small areas first. They'll have to get over regulatory and law hurdles, but this can be coming sooner than later.

If it's 30 years out, then it won't matter to many of us.

I don't think these companies are expecting that long a period for turnover.

I think they do for a TOTAL turn over. But think differently. I am starting to get it. There will be SDC areas or zones, sooner than later. Maybe for commercial purposes.

Everyone buying an SDC to travel cross country is likely farther away.

 

[JoeWras]

Yes, a maker shows what its car can do in a demo in a nice clean scenario. He is not going to show what his car cannot do. It's common sense.

NW and I have both been on the other side of engineering demos. It is messy business and I think is part of what is causing us to tap the brakes on this.

Engineers have been demoing optimistic stuff since the time of the Pharaohs. I doubt SDCs are any different.

 

[Walt34]

Fortunately 2015 (latest) data shows 28% of auto crash fatalities were unrestrained occupants, so we've gotten much better? Or air bags have helped?

I'll never forget the first time in the late '70's when I walked up to the scene of a single-car accident where the air bags had deployed. I thought for sure it was going to be a fatal or near-fatal. Nope - just "bumps 'n bruises" on the driver. Made a believer out of me.

Back then air bags were an expensive option. That guy got his money's worth.

 

[NW-Bound]

NW and I have both on the other side of engineering demos. It is messy business and I think is part of what is causing us to tap the brakes on this.

Engineers have been demoing optimistic stuff since the time of the Pharaohs. I doubt SDCs are any different.

And at megacorp, I along with others were fooled by a subcontractor. His demo worked well for a prototype, and just needed some tweaks it looked like. And he showed a better understanding of this technology than his competitors. Lots of PhD's working the program. So, he got the contract. Remember that in aerospace, quite often it is not so much about costs as a guarantee that it is going to work. The ultimate customer is willing to pay big bucks.

Oh man, in preproduction testing, his system showed weird results that he did not predict. At this point, we had to work with him to solve the problem as we were in it deep too. We tested, and checked, and tested, and checked. Random results all over the place. Too late to go to another supplier. What a fiasco! I left that program when it was still going on. Have not bothered to check with people I knew to see what the end result was.

PS. As usual, at megacorp there were different groups backing different vendors and technology. When things did not work out, there was a lot of "I told you so" and fingerpointing. Many projects came to a crashing halt, literally (things crashed and burned up bad). Very scary. Very sad.

PPS. I could tell you some keywords, then you can go search the Web for these "fantastic" failures after the corp spending mucho of its own money. Aviation Week (we call it Aviation Leak) knows about this stuff. Their reporters really snoop around.

 

[Midpack]

NW and I have both been on the other side of engineering demos. It is messy business and I think is part of what is causing us to tap the brakes on this.

Engineers have been demoing optimistic stuff since the time of the Pharaohs. I doubt SDCs are any different.

As I noted earlier, I was too, like others here. As an engineer I interfaced with Sales/Marketing, daily for the last 18 years, and every other corp discipline including the suits during my career too. I am all too familiar with Marketing overpromising and leaving me & my group to make it happen.

Like almost everyone at ER.org, I just don't need to flash my "credentials" every third post hoping it'll add weight to my POV (not you JW). Like most here, I try to refrain from sharing my background. There are scads of accomplished professionals here from all fields. Truly, some of the smartest most interesting people I've known had nothing but a high school degree. And some of the dumbest (in practical terms) and dullest, with little common sense, had postgraduate degrees. Let each POV stand on its own merits.

 

[JoeWras]

I just don't need to flash my "credentials" every third post hoping it'll add weight to my POV (not you JW). Let each POV stand on its own merits.

Ha ha, I don't have any credentials in SDC. And I know you were not poking at me. The software they do is far removed from mine. The sensors are a different world.

I only have the experience of getting pressured from management to show something half-baked. I've scraped off a lot of burnt crust and shown beautiful bread many times. This gives me pause.

 

[Ronstar]

..There will be SDC areas or zones, sooner than later. Maybe for commercial purposes.



Everyone buying an SDC to travel cross country is likely farther away.

I agree. I think this will happen very soon in private settings. Maybe industrial parks where self driving vehicles transport people, parts, equipment between multiple buildings/ facilities.

 

[NW-Bound]

...Like almost everyone at ER.org, I just don't need to flash my "credentials" every third post hoping it'll add weight to my POV (not you JW). Let each POV stand on its own merits.

We dealt with a lot of technology where the vendors or subcontractors knew far more about their stuff than we did. That's why we could not do it ourself and needed them. Duh! But people have learned to be skeptical and asked a lot of questions.

I am sure that Honda and Ford have asked Waymo a lot of questions and are shown a lot more than the public is shown. And so, I will just have to accept their schedule of introduction to the public. It's not that far away (I am impressed), so we will see. I do not see what keeps them from a level-4 car at this point, so my ponderings are more about cost, if you notice.

No, it is not "credentials" that I flashed. I was just retelling real-life experiences where we got unexpected results from dealing with technology that we ourselves were laymen. This fancy high-tech stuff can be tricky.

My son who worked in the micro-electronic stuff (that goes in your cell phone) told me of things I was totally oblivious about. Not my field at all. Then it dawned on me why these things do not last as long as expected. Solid-state chips are supposed to last much longer than that (I will not describe them here though. Got into enough hot water. ).

 

[ERD50]

...

No, it is not "credentials" that I flashed. I was just retelling real-life experiences where we got unexpected results from dealing with technology that we ourselves were laymen. This fancy high-tech stuff can be tricky. ....

I never saw it as 'flashing credentials', just explaining some background and relevant experience. Many posters here fill us in on their background, so we understand where they are coming from. Just off-hand, one poster has experience in fraud investigations, another as a supplier to big auto companies, some with medical backgrounds, a surveyor,lumber dealer, etc. It's interesting to hear some of the inside baseball.

-ERD50

 

[NW-Bound]

Thanks.

After I mentioned Aviation Week as a trade magazine for people in aerospace, it occurred to me that the auto industry would have many trade magazines that cover this technology. It would be of interest to people who want the latest scoop. I have enough of this stuff, just from this thread.

As for me, the promised launch of these cars is close enough, and only a few years away. I can wait to see what companies offer, and what the price is. If they are better than I expected, I am happy. I may get to buy a car I will enjoy. It may be time for me to get a new car in a few years too.

The chart Midpack posted shows a Honda capable of driverless on the highway in 2020. I think it means it has level-4 capability, but on the highway, not in the city. It's OK. It means a super reliable and safe highway cruiser, which nobody now has. But hopefully the price is reasonable. Not sure if I will have any money left after paying for healthcare insurance, as I am still not on Medicare in 2020. That's more my concern.

 

[NW-Bound]

I think I should point out something again, in case it is lost. It is that one does not have to be an expert in this technology to notice something. He just has to pay attention.

That is, as I noted in an earlier post, watching the video by Waymo CEO that Midpack posted, the new super Waymo sensor suite employs all 3 sensor technologies: lidar, radar, and vision camera. And all 3 have 360-degree coverage. They said they needed it for a level-4 car, and eventually level 5.

Without knowing much about what each of these sensors can do, one will immediately ask: if the leader in this technology said he needed that much, then how does a company with fewer sensors solve the same problem?

If you can provide the same capability with fewer sensors, then you are smarter than the leader. It's up to you to demonstrate that.

I am not an expert in this field at all. I observe, compare, and ask questions.

 

[Midpack]

I never saw it as 'flashing credentials', just explaining some background and relevant experience. Many posters here fill us in on their background, so we understand where they are coming from. Just off-hand, one poster has experience in fraud investigations, another as a supplier to big auto companies, some with medical backgrounds, a surveyor,lumber dealer, etc. It's interesting to hear some of the inside baseball.

It's not flashing if it's done sparingly. But when it's repeated dozens of times in a thread it strongly suggests another purpose.

 

[Ronstar]

There is one bit of new gps technology that could hasten the development of SDC's. The gps block III satellites are scheduled to begin launch in 2018. They have at least one extra civilian signal and have submeter accuracy. Certainly an improvement over current gps accuracy.