2019 CAR Management Briefing Seminar talk level 5 Autonomus Driving

[Miko M]

Interested article about level 5 autonomous driving. Some Automakers won't like this but I am sure most of us love it.

https://carbuzz.com/news/this-is-not...s-want-to-hear

This Is Not What Some Automakers Want To Hear

Could this be a problem in the very near future?

The dream the public is being sold is that in just a few years we will all be sitting in cars, such as the upcoming Tesla Model Y, with no steering wheels that drive us to and from our destinations with no driver input other than a destination. There will be no more crashes, no more pedestrian fatalities, no more traffic jams, no more hunting for parking spaces, and we'll use all that to get more things done in a day. Life will be easy as we drink our coffee and watch Netflix on the way to work, and nobody need ever get a traffic ticket again.
The reality is that a lot of new and old companies have put their eggs in that basket and, according to a report from Automotive News, the portfolio director of safety domain control units for the German parts supplier ZF, Farid Khairallah, is publicly saying that the reality of that dream may be out of reach. ZF is a company famous for supplying transmissions to many automakers but has also put a lot of money into self-driving cars.

The problem gets exponentially more complicated the closer the systems get to 100% autonomy, and complicated means expensive. For an autonomous vehicle to be 100% safe 100% of the time, the computer power required to assess information and process bank of cameras, radars, and lidars at the speed necessary is immense. Then Khairallah points out that those systems need cooling systems as well, and support systems crank the cost even more.
"The whole industry is rethinking their strategies and what they want to do with this," Khairallah said.
We may not find ourselves riding around in autonomous vehicles in our lifetimes, but that doesn't mean it won't be a reality one day. It's not all bad news though. So far we've seen some huge benefits from Level 2 development such as automatic emergency braking and lane-keep assist become standard features for many car models.

[M 4EVER]

You are posting good articles MIKO. Thank you for sharing

[Poochie]

Yes, just like everything else that came before, it's initially expensive but as the technology evolve, the cost would surely drop..

That article is just click-bait and they're not saying anything we didn't already know..

In 2006, Lexus was the first company to implement a self-parking feature in their $100,000 LS 460..

We all thought it was revolutionary and would take centuries for that technology to trickle down to the masses but fast forward 10 years later, it comes standard in an average Kia.

Give it time, just like everything else, the cost of will become reasonable. Tesla already hinting at this with their reasonable-priced Model 3..

I put on my tin foil hat for nothing - next


.

[Alfisti]

I called it years ago, the level of complexity in absolutely insane. The ONLY way it works is to ban driving and go full autonomous where EVERY car on the road talks to each other. Otherwise this has been a colossal waste of money, stunned as to the gullibility of leading auto makers to throw cash at this insanity. it was always out of reach.

[clee1982]

if wireless network is reliable enough you can have the "hard" computational going on outside car, but that will take a while I suppose (and as of today I guess communication overhead still cost more than computational overhead though that might change)

[Efthreeoh]

Pretty picture; any half-competent driver can drive in that environment.

Change to reality: rain, snow, high winds, failing hardware, software bugs, tailgating RAM pickup trucks, animals in the road....

Yeah. I'll wait.

[Efthreeoh]

Originally Posted by clee1982

if wireless network is reliable enough you can have the "hard" computational going on outside car, but that will take a while I suppose (and as of today I guess communication overhead still cost more than computational overhead though that might change)

Air traffic management is highly automated, but the end, the computational activity is... yup, air traffic controllers. Human brains...

[clee1982]

I don't doubt you and don't think it's an easy problem to solve, bit even 2 decade away is "very soon", 2 decade means all my kids won't be driving themselves. That's a lot faster than horse wagon to cars.


https://thetyee.ca/News/2013/03/06/H...ung-Big-Shift/

Edit: you can ignore the political forgone conclusion part but interesting history to read...

[Efthreeoh]

Originally Posted by clee1982

I don't doubt you and don't think it's an easy problem to solve, bit even 2 decade away is "very soon", 2 decade means all my kids won't be driving themselves. That's a lot faster than horse wagon to cars.


https://thetyee.ca/News/2013/03/06/H...ung-Big-Shift/

Edit: you can ignore the political forgone conclusion part but interesting history to read...

The political push to EV is captured perfectly in this excerpt from the piece:

"One hundred years later, North Americans still hang onto their stubborn beliefs that there is a quick technological fix to every energy conundrum, says Greene."

Your statement about 20 years to autonomy will be faster than the full transition from horse to automobile is a bit misleading. It's not the same type of transition, and it doesn't account for the economic expansion the automobile brought.

[clee1982]

I'm not trying to make an analogy on any of the policy implication, all I want to say is it took 50 year to go from horse to car, so I can see it happen in 20 years to full autonomous driving, and that is pretty quick by my book

[Efthreeoh]

Originally Posted by clee1982

I'm not trying to make an analogy on any of the policy implication, all I want to say is it took 50 year to go from horse to car, so I can see it happen in 20 years to full autonomous driving, and that is pretty quick by my book

The only real take away from that article is... converting energy to do work creates "exhaust". He turned it into a political discussion hidden within some stupid history lesson that may or may not be accurate; but more, it really missed the point regardless. It didn't take 50 years to convert from horses to cars.

[clee1982]

Ah I was just trying to find an evidence to back up my claim of taking 50 years from horse to car..., don't care for the politics

[Efthreeoh]

Originally Posted by clee1982

Ah I was just trying to find an evidence to back up my claim of taking 50 years from horse to car..., don't care for the politics

At the age of 7 in 1935, my mother travelled across the US from Norfolk, VA to San Francisco in a 1932 Chevrolet 4-door sedan with her parents and two younger brothers. The American automobile industry was started in earnest in the early 1900's, really between 1903 - 1908. So in less than 30 years the US was able to create an entire new transportation system that allowed motorized travel across the 3,000 mile country. That involved creating a paved road system, fuel delivery infrastructure, and fuel stops and hotels plentiful enough for such safe travel. Twenty years to possibly get to Level 5 autonomy, let alone complete dissolution of human-driven cars, based on the speed of technology advancement in the modern era, I'd say is quite slow compared to what effort was undertaken to create the automotive transportation system in the early 20th century.

If the author of that article is counting the start at the invention of the automobile in the mid 1880's then sure, it was near 50 years, but from a realistic historical perspective that is a bit disingenuous.

[Efthreeoh]

Originally Posted by Alfisti

I called it years ago, the level of complexity in absolutely insane. The ONLY way it works is to ban driving and go full autonomous where EVERY car on the road talks to each other. Otherwise this has been a colossal waste of money, stunned as to the gullibility of leading auto makers to throw cash at this insanity. it was always out of reach.

Exactly. Converting the national fleet of personally-owned automobiles to Level 5 (no steering wheel or pedals) overnight just is not going to happen. The only politicians that have the balls to suggest the Government can mandate to take personally-driven cars away are AOC and Bernie, both who are idiots...

[RM7]

Given our refusal to actually crack down on distracted driving by mandating phone companies turn them off above 5mph or something similar, autonomous driving can't come fast enough. I like driving as much as the next guy on this site, but it's becoming more dangerous for everyone and we refuse to "give up any rights", such as the right to text and check facebook while driving. This isn't a generational thing, I see drivers of all ages doing it every day.

[Efthreeoh]

Originally Posted by JamesNoBrakes

Given our refusal to actually crack down on distracted driving by mandating phone companies turn them off above 5mph or something similar, autonomous driving can't come fast enough. I like driving as much as the next guy on this site, but it's becoming more dangerous for everyone and we refuse to "give up any rights", such as the right to text and check facebook while driving. This isn't a generational thing, I see drivers of all ages doing it every day.

The companies behind autonomous are Big Data, so they'll never allow a mandate to disallow use of phones while driving, nor self police it. They love to track you where ever you go. The Government would just love to implement autonomous driving so you have to log your trip and then you can be constantly tracked and controlled. Kiss your privacy goodbye (even more than it is already gone).

[RM7]

Originally Posted by Efthreeoh

The companies behind autonomous are Big Data, so they'll never allow a mandate to disallow use of phones while driving, nor self police it. They love to track you where ever you go. The Government would just love to implement autonomous driving so you have to log your trip and then you can be constantly tracked and controlled. Kiss your privacy goodbye (even more than it is already gone).

The government could care less about tracking you and where you go relative to highway safety, the agencies like DOT, NHTSA and NTSB want to reduce the deaths on the road. Congress controls the ability to make and pass laws and congress is controlled by money and people. Too many people are convinced "they" can handle a phone while they drive and it's the "other people" that should be somehow policed. It'll take the Presdient's daughter getting creamed by a distracted driver most likely to do something about this. Until then, the deaths and injuries will continue to stack up until we have autonomous driving. Not that it will be without challenges and hiccups.

[Efthreeoh]

Originally Posted by JamesNoBrakes

The government could care less about tracking you and where you go relative to highway safety, the agencies like DOT, NHTSA and NTSB want to reduce the deaths on the road. Congress controls the ability to make and pass laws and congress is controlled by money and people. Too many people are convinced "they" can handle a phone while they drive and it's the "other people" that should be somehow policed. It'll take the Presdient's daughter getting creamed by a distracted driver most likely to do something about this. Until then, the deaths and injuries will continue to stack up until we have autonomous driving. Not that it will be without challenges and hiccups.

I have professional experience is this area. The Congress and the DOT have no idea how to implement an autonomous driving transportation system. The only way an autonomous transportation system will lower death rates is by reducing overall trips and lowering overall road speeds to lessen crash impact energy.

[RM7]

Originally Posted by Efthreeoh

I have professional experience is this area. The Congress and the DOT have no idea how to implement an autonomous driving transportation system. The only way an autonomous transportation system will lower death rates is by reducing overall trips and lowering overall road speeds to lessen crash impact energy.

So do I, and I call BS on that, sorry. I didn't say they know how to implement autonomous driving right now, but they and we will get there. I said they want to reduce traffic deaths. These issues are being worked on currently. By metering against the vehicle in front, keeping distances, slowing down when the vehicle ahead slows down (rather than having the head down on the phone), staying in the lane, braking for someone that suddenly crosses (when again you might be on your phone) and a plethora of other situations, it will absolutely reduce deaths. Other motorists. Pedestrians. You name it. That doesn't mean lowering overall speeds. It may mean averaging speeds.

[Efthreeoh]

Originally Posted by JamesNoBrakes

So do I, and I call BS on that, sorry. I didn't say they know how to implement autonomous driving right now, but they and we will get there. I said they want to reduce traffic deaths. These issues are being worked on currently. By metering against the vehicle in front, keeping distances, slowing down when the vehicle ahead slows down (rather than having the head down on the phone), staying in the lane, braking for someone that suddenly crosses (when again you might be on your phone) and a plethora of other situations, it will absolutely reduce deaths. Other motorists. Pedestrians. You name it. That doesn't mean lowering overall speeds. It may mean averaging speeds.

Call BS if you want, but a review of legislation on the subject of AV shows Congress has no comprehensive plan for implementation of a Level 5 system. Cars equipped with sensors and AI are not going to solve congestion. Congestion leads to phone use. Congress at best has tried to address a national standard for regulating autonomous vehicles, rather than develop regulation and standards for an "AV system". That's only to fend off state-level AV legislation. So outside of herding cats, Congress has no comprehensive plan. When the public learns what the true cost of converting to a Level 5 AV transportation system, both at the taxation level and vehicle cost level, its desire to convert to AV may wane.

Any AV system will need to lower impact speeds to lower death rates. Collisions are unavoidable as long as cars drive on streets with cross traffic, regardless if humans or computers are driving the vehicle.

[RM7]

Originally Posted by Efthreeoh

Call BS if you want, but a review of legislation on the subject of AV shows Congress has no comprehensive plan for implementation of a Level 5 system. Cars equipped with sensors and AI are not going to solve congestion. Congestion leads to phone use. Congress at best has tried to address a national standard for regulating autonomous vehicles, rather than develop regulation and standards for an "AV system". That's only to fend off state-level AV legislation. So outside of herding cats, Congress has no comprehensive plan. When the public learns what the true cost of converting to a Level 5 AV transportation system, both at the taxation level and vehicle cost level, its desire to convert to AV may wane.

Any AV system will need to lower impact speeds to lower death rates. Collisions are unavoidable as long as cars drive on streets with cross traffic, regardless if humans or computers are driving the vehicle.

That doesn't quite make sense. Congress would tell the DOT to prepare and make rules, they'd call in their experts and engineers, get advisory boards with other experts, and draft rules. Congress usually doesn't make these kind of rules or draft plans except in rare cases and usually when they do, they have to "fix" them for the exact reason you explain, because they aren't experts. They only set the more general policy/rules. The "how to do it" falls on the agencies.

Secondly, you are the only one saying it's going to solve congestion. Obviously it's not, you need more capacity or to have more passengers if you are going to solve congestion. That's a different issue. Safety is the autonomous driving issue. Again, most lawmakers and and agencies are concerned with deaths and injuries. Concerns about congestion usually far pretty far below this. How does congestion lead to phone use? If you are taking stop-and-go traffic, well that probably limits deaths/injuries just by the nature of slow speeds, but that's a heck of a stretch. Does congestion lead to eating food in the car too? Dealing with kids? Talking on the phone? The idea that you aren't going to have any congestion, no stop-n-go, no stop-signs or lights seems pretty far fetched given city density and lack of public transportation. Is that really where the most bang for the buck is? You can only make highways so wide and there will still be choke points and merges that you can't make better. Just look at some of the mega-on-ramp-overpass complexes in Texas. My god, it's just getting out of control in some of those places, almost better to start from scratch.

I disagree that the only way to save lives or have less injuries is to slow speeds. Keeping cars from careening off the road is another way. Keeping cars from hitting other cars, like on-coming traffic, is yet another, keeping them from slamming into a stationary car at 70mph is yet another, and so on. These rely on the autonomous technology. Lowering deaths by lowering speeds to have less injuries and damage in crashes is not the entire principle of this, as you claim. Autonomous cars can be zipping along at 70mph and still be safer, because they won't drift off, hit an on-coming car, run over a motorcycle, etc.

[Efthreeoh]

Originally Posted by JamesNoBrakes

That doesn't quite make sense. Congress would tell the DOT to prepare and make rules, they'd call in their experts and engineers, get advisory boards with other experts, and draft rules. Congress usually doesn't make these kind of rules or draft plans except in rare cases and usually when they do, they have to "fix" them for the exact reason you explain, because they aren't experts. They only set the more general policy/rules. The "how to do it" falls on the agencies.

Secondly, you are the only one saying it's going to solve congestion. Obviously it's not, you need more capacity or to have more passengers if you are going to solve congestion. That's a different issue. Safety is the autonomous driving issue. Again, most lawmakers and and agencies are concerned with deaths and injuries. Concerns about congestion usually far pretty far below this. How does congestion lead to phone use? If you are taking stop-and-go traffic, well that probably limits deaths/injuries just by the nature of slow speeds, but that's a heck of a stretch. Does congestion lead to eating food in the car too? Dealing with kids? Talking on the phone? The idea that you aren't going to have any congestion, no stop-n-go, no stop-signs or lights seems pretty far fetched given city density and lack of public transportation. Is that really where the most bang for the buck is? You can only make highways so wide and there will still be choke points and merges that you can't make better. Just look at some of the mega-on-ramp-overpass complexes in Texas. My god, it's just getting out of control in some of those places, almost better to start from scratch.

I disagree that the only way to save lives or have less injuries is to slow speeds. Keeping cars from careening off the road is another way. Keeping cars from hitting other cars, like on-coming traffic, is yet another, keeping them from slamming into a stationary car at 70mph is yet another, and so on. These rely on the autonomous technology. Lowering deaths by lowering speeds to have less injuries and damage in crashes is not the entire principle of this, as you claim. Autonomous cars can be zipping along at 70mph and still be safer, because they won't drift off, hit an on-coming car, run over a motorcycle, etc.

Congress has to set the goals and funding levels; that is what legislation does. The agencies then try to achieve those goals within the time frame and budget the legislation states.

I drive on Route 66 outside of Washington DC. Between Manassass, VA and Centerville, VA 66 backs up to a crawl most days. I see a lot of drivers on their phones more when the traffic slows than when it is at the 60 MPH posted speed limit. Then the distracted phone use exacerbates the traffic congestion because drivers are not fully paying attention to the traffic flow, which means they don't speed up when they should, or they stay 8 car lengths behind the vehicle in front of them because they are finishing a text. In your vision, AV would solve this problem of distracted driving thus reducing congestion. But as you stated, and I wrote earlier in this thread, congestion is a result of cross traffic at intersections, where the road capacity is overwhelmed and can't store enough vehicles waiting at intersections, so the traffic backs up on other primary roads.

The only way AV can solve that situation is to control movement of each car and keep two cars out of the same spot on the road at the same time. I do not think the car-to-car AV systems now and even in the future can prevent all collisions in that scenario. Software and computers are not near perfect (despite what engineers will tell you - I work with plenty of them ), and at the speeds involved, redundant backup systems are probably not fast enough to recover in case of a vehicle equipment, hardware or software failure. The short distance of vehicle separation at high speed has been one of the touted attributes of a Level 5 automated traffic system, suggested as reducing congestion because more cars can be safely packed into a given roadway space (at higher speed), as it is considered computers are better drivers than humans. So, it not me making those assertations, it's the AV industry.

Again, when cars back up onto major roads from secondary roads because throughput at intersections are the cause, speeds resultantly slow on the primary roads whether a human or computer is driving. Timing at intersections will take route planning by a local master transportation control system and over-the-air comms to each vehicle. All those systems are in use today in the air traffic control industry and none are infallible. The speed, traffic density, and two-dimensional space make the automated separation cars way harder than airplanes.

I'm not saying the engineering challenges can't be overcome, but I am saying the cost to truly make a safe driving environment via autonomous vehicles is going to be incredibly expensive. My opinion is the side effects may not be liked by today's public (but maybe tech-wheenie, government controlled future citizens may think differently). When the Government controls the flow of traffic, which it eventually will have to, when the system gets overwhelmed, and it will, in order to maintain levels of safety (stated as the main goal of the DOT, NTSB, and NHTSA), the Government will curtail access to the system, which means it will prevent citizens from travelling when and where then want to, and slow the rate of traffic. Today's public will not tolerate such a system.

And you mentioned motorcycles...I'd love to know how they fit into the system. And this all gets migrated to level 5 while humans still get to drive their old-fashion oil burners... "First, kill all the lawyers"...