Will we see a real self driving car in our life time?

nsaspook

Joined Aug 27, 2009
13,079
Woz was always the smartest guy at Apple.

The weirdest thing is everything/everyone is in a denial mode but the technology is still being worked upon. Its becoming a rat race.
SDC's are going to work, they might just not roll out for the general public soon enough. The infrastructure keeps changing, its keeps remodeling itself with time and technology. Problem is we are not sure what we really should do, since SDC's are pretty much in research phase.
People like him who understand technology down to the chip level know about the 80/20% rule. The last 10% is what makes or breaks computerized products. It seems relatively easy to be better than a human 90+% of the time when driving is boring and routine. This means that stupid accidents from distraction or lapses in simple driving rule judgement will be reduced. That last few % involves novel events with infinite variations you can't train with a computer dataset. It's something even low skilled humans with intelligence can mainly handle without any specific training because of a millions years of brain evolution that allow us to construct the likely future of a series of events in order to survive in the face of danger. I can see one day getting to 99% if we greatly reduce the human interaction with 'AI' cars but I see absolutely no reason give up driving my own car even if machines were perfect because I enjoy driving as a daily adventure.

https://www.quora.com/Engineering-M...-of-2-3/answer/Michael-Wolfe?srid=24b&share=1
 

marcuskeene

Joined Oct 15, 2018
27
I can see one day getting to 99% if we greatly reduce the human interaction with 'AI' cars but I see absolutely no reason give up driving my own car even if machines were perfect because I enjoy driving as a daily adventure.
And you can continue doing that. Self driving cars wont make human driving obsolete. Simple. But it could be really helpful to those who 'can't'. To assist people who can't function like you or me, it would be a great asset. There's more to technology than just disruption and plucking out the human.
 

nsaspook

Joined Aug 27, 2009
13,079
And you can continue doing that. Self driving cars wont make human driving obsolete.
Making human driving obsolete is the end game of the technology and most people in the AI field believe that.

https://www.theverge.com/transportation/2015/3/17/8232187/elon-musk-human-drivers-are-dangerous
Tesla co-founder and CEO Elon Musk believes that cars you can control will eventually be outlawed in favor of ones that are controlled by robots. The simple explanation: Musk believes computers will do a much better job than us to the point where, statistically, humans would be a liability on roadways.

"I don't think we have to worry about autonomous cars, because that's sort of like a narrow form of AI," Musk told NVidia co-founder and CEO Jen-Hsun Huang at the technology company's annual developers conference today. "It would be like an elevator. They used to have elevator operators, and then we developed some simple circuitry to have elevators just automatically come to the floor that you're at ... the car is going to be just like that." So what happens when we get there? Musk said that the obvious move is to outlaw driving cars. "It's too dangerous," Musk said. "You can't have a person driving a two-ton death machine."
...
Today, Musk noted that the hardest part of helping cars drive themselves is what happens when vehicles are traveling between 15 and 50 miles per hour. "That's where you get a lot of unexpected things," Musk said. That list includes road closures, open manhole covers, children playing, and bicyclists. Lots of things that your robot car could run into without human remorse.
 

Alec_t

Joined Sep 17, 2013
14,280
There may be a gender issue when it comes to self-driving cars. Male drivers tend to want to be 'in control' of their cars and so would be less likely than women to accept self-drive. Do you see self-driving cars having a 'male/female' switch on the dashboard to select drive mode?
And what is the future for high performance sports cars? Effectively banned? Safety rules and regulations built into the algorithms could be so restrictive that all cars would be constrained to some bog-standard sedate performance.
Problems, too, to look forward to when you take your left-hand-drive car abroad to a country where right-hand-drive is the norm!
 

shortbus

Joined Sep 30, 2009
10,045
Self driving cars wont make human driving obsolete.
But to be successful that is the only way it will happen in the end. While AI or what ever can handle the hardware side of things(they pretty much can do it now), The software side is a none starter until humans doing their thing while driving is the problem. If only self driving cars are the only ones on the road, the software side is then done between each vehicle talking to each other.
 

marcuskeene

Joined Oct 15, 2018
27
But to be successful that is the only way it will happen in the end. While AI or what ever can handle the hardware side of things(they pretty much can do it now), The software side is a none starter until humans doing their thing while driving is the problem. If only self driving cars are the only ones on the road, the software side is then done between each vehicle talking to each other.
It sometimes is confusing for me to understand, what exactly is the issue? Self-driving cars can't be trusted? or They will take over human drivers?

These are two different matters altogether and no I don't think that major challenge is the human driving (since most of these cars would still recommend an alert driver on the steer) but the infrastructure.
I remember someone once pointing out that it doesn't matter how slow the SDC is? If there are kids running on the road, its bound to result in an accident.
With this sort of infrastructure, no matter who/what drives the car, you can never expect the mishap ratio coming down.
 

CarExtreme

Joined Oct 12, 2018
4
The weirdest thing is everything/everyone is in a denial mode but the technology is still being worked upon. Its becoming a rat race.
SDC's are going to work, they might just not roll out for the general public soon enough. The infrastructure keeps changing, its keeps remodeling itself with time and technology. Problem is we are not sure what we really should do, since SDC's are pretty much in research phase.
You’re right. It’s too early to tell what kind of infrastructure SDCs will have. But I guess the transition has begun with Waymo’s self-driving service launching next month. That literally, could be the “test-drive” for this future technology.
 

Glenn Holland

Joined Dec 26, 2014
703
You’re right. It’s too early to tell what kind of infrastructure SDCs will have. But I guess the transition has begun with Waymo’s self-driving service launching next month. That literally, could be the “test-drive” for this future technology.
For the past 4 years, I've been reading about how Joe Schmoe SDC Inc. will roll out self driving cars next year. To date, I still haven't seen one of them in common use.
 

marcuskeene

Joined Oct 15, 2018
27
You’re right. It’s too early to tell what kind of infrastructure SDCs will have. But I guess the transition has begun with Waymo’s self-driving service launching next month. That literally, could be the “test-drive” for this future technology.
I have no issues with it being delayed as long as it delivers its objective. We must learn from what went wrong previously.
This is the latest on public testing by Daimler and Bosch.
https://www.topspeed.com/cars/car-n...-hailing-service-in-san-jose-ca-ar183389.html
 

CarExtreme

Joined Oct 12, 2018
4
I have no issues with it being delayed as long as it delivers its objective. We must learn from what went wrong previously.
This is the latest on public testing by Daimler and Bosch.
https://www.topspeed.com/cars/car-n...-hailing-service-in-san-jose-ca-ar183389.html

Delay is okay, I just hope the government would be more supportive though. I’m baffled as to why some states are putting the brakes on testing. Case in point: In NYC, General Motors Co. has put on hold plans to begin testing in Manhattan because Mayor Bill de Blasio has expressed concerns about the technology’s safety.

Hello? That’s why we need a testing phase.
 

Glenn Holland

Joined Dec 26, 2014
703
Delay is okay, I just hope the government would be more supportive though. I’m baffled as to why some states are putting the brakes on testing. Case in point: In NYC, General Motors Co. has put on hold plans to begin testing in Manhattan because Mayor Bill de Blasio has expressed concerns about the technology’s safety.

Hello? That’s why we need a testing phase.
Sounds like the congress's excuse for passing Obama Care: "We had to pass it to know what's in it".

Using the same flawed reasoning about SDCs, "We need to let them get in accidents to know if they are unsafe".
 

marcuskeene

Joined Oct 15, 2018
27
Sounds like the congress's excuse for passing Obama Care: "We had to pass it to know what's in it".

Using the same flawed reasoning about SDCs, "We need to let them get in accidents to know if they are unsafe".
By that logic, we must not get on air planes, they crash.
I am sure those testing these SDC's know their role and responsibility. We can't make past incidents a rule. Not that technology doesn't progress/improve. I am sure they have thought it through too.
 

Glenn Holland

Joined Dec 26, 2014
703
Aircraft are proven technology and not some pie in the sky venture that exists only because the defense department (DARPA) is handing out big $$$ to fund it.
 

nsaspook

Joined Aug 27, 2009
13,079
I am sure, people back in the day felt the same about aircraft too.
It's more like marketing supersonic aircraft for the general public. We could do it, we built it and it flopped cause it didn't deliver value to the general public. Simply because it's possible after a huge amount of investment doesn't make it 'the next big thing'. No matter how good the technology is, success won’t happen without public support.
Efforts to educate the public.
https://www.bosch-mobility-solutions.us/us/highlights/automated-mobility/amc/

There is a huge amount of inertia bound into regular humans driving Level 0 cars for over a hundred years embedded into the driving infrastructure.
https://www.fool.com/investing/2018/09/06/self-driving-cars-understanding-6-autonomous-level.aspx
But note that Waymo and GM aren't choosing locations at random. Waymo spent months "training" its prototype vehicles on the roads and traffic conditions in and around the Chandler area; GM Cruise has been doing the same with its Chevrolet Bolt-based vehicles in San Francisco. For now, those vehicles are limited to the areas with which their systems are most familiar, all of which have been very carefully mapped.

Those limits are what make the systems Level 4. For instance, no matter how the maps are used, if the vehicle is dependent on a map, that means it's limited, because it can't go anywhere that isn't yet mapped. Of course, a Level 4 system could have other limits; it may not switch on if it detects heavy snow, for instance.

Long story short: If a system offers full automated driving within limits, the limits are what makes it Level 4.
Why full self-driving might still be years away
Full self-driving is an incredibly difficult problem to solve. Consider that even Waymo -- which began way back in 2009 as the Google Self-Driving Car Project and has some of the best and most experienced engineers in the field -- is still having trouble getting its vehicles to safely navigate some routine traffic situations in suburban Chandler, where the situations are all known and the area is carefully mapped. How do you think they'd fare in a completely unfamiliar environment -- downtown Mumbai, for instance?
Self-driving cars MUST be totally compatible with that to have any hope of developing beyond a boutique Level 4 (prescribed areas, or on certain routes) product in specialized areas with low consumer acceptance ratios. Unless we have real Level 5 (can drive in all conditions) autonomous vehicles instead of mainly Level 3/4 (capable of full control IF conditions are met) autonomous vehicles in the beginning the results won't be pretty given the expectations of the general public to do other activities or sleep while the car is driving anywhere.

What's the value (in the existing driving infrastructure) of complete self-driving cars to the general public over just advanced driver assist (Level 1/2, keep the car in its lane and a safe distance from other cars) with a human always in control to eliminate handoff problems?
 

nsaspook

Joined Aug 27, 2009
13,079
https://www.cnn.com/2018/11/21/success/waymo-ceo-driverless-cars/index.html
We sat outside with a couple of sodas and a box of fries and talked about what automated cars, even with all those sensors and all that experience, won't be able to do. Engineers talk about L5, or Level 5 autonomous driving. That's as autonomous as any vehicle could possibly be. At L5, you could get into a car at any time under any weather conditions and say "Take me...." anywhere. But L5 is impossible, said Krafcik.
"That's a somewhat unreasonable expectation," he said, "and it's sort of not even necessary because we so rarely drive from San Francisco to Santiago, Chile. We just don't do that."
...
Waymo cars will operate in designated areas on familiar roads. Those designated areas will widen over time covering more places. But Krafcik doesn't foresee a day when the need for human drivers will be entirely obviated. But that need will diminish.
Even then, though, no one will take the steering wheel from Krafcik's hands. Not even his own company's creations.
 

Glenn Holland

Joined Dec 26, 2014
703
When it comes to control system engineering, I'm obviously not the sharpest knife in the drawer. However I'm wondering if Google/Waymo, GM/Cruise, etc. are familiar with so called "Positive Train Control" (PTC) used in the railroad industry and mass transit?

PTC is based on a driving algorithm applied to rail vehicles and it uses acceleration and speed sensors to calculate the safe following distance to the train ahead. However, it is a networked system where position and speed data of each individual train is transmitted to a central computer and the maximum allowable speed is transmitted back to the train. In addition to speed and acceleration sensors, there are also checked redundant features such as axle counters and cross over loops along the trackway to monitor the approximate position of each train.

In order for self driving cars to operate to an extremely high level of safety, the control system must be networked and checked redundant and I don't believe that any of the companies are using this technique.
 
Top