Hyundai IONIQ Forum banner
1 - 20 of 52 Posts

·
Registered
Joined
·
1,443 Posts
Discussion Starter · #1 ·
Yes, it's lazy and takes away the joy of driving but on certain journeys.. I would love the option of letting it drive for me. I have half the feeling of autonomous driving with the Ioniq's ACC but I sometimes find myself wanting it to take the wheel as well.

So hyundai's working on one and Tesla and probably a few others but I don't know the details, will it be an optional feature or a car designed purely for this type of driving?

Also, what are people's views on autonomous cars? Safety is a big thing, we've seen when tech fails.. it's not pretty but I'm sure overrides would be put in place for such times.

Discuss..
 

·
Registered
Joined
·
260 Posts
The big issue with autonomous cars for me is the choice it will have to make at some point between life and death if either the occupant of the car itself, or what it is about to plough in to.

I am all for systems like LKAS, AEB and ACC, but all these should be seen as is aids to the driver. With AEB yes it's great for those moments where you don't concentrate, but if you are doing 60 on icy roads, it's unlikely that will save you from a shunt.

Autonomous cars are fine in principal, but there are many things I think why they shouldn't be used on roads where other users of 'non' intelligent cars are.
 

·
Registered
Joined
·
162 Posts
The big issue with autonomous cars for me is the choice it will have to make at some point between life and death if either the occupant of the car itself, or what it is about to plough in to.
Humans are already making that choice, far more often than needed & usually end up making a bad one. The number of people who die every year in car accidents is staggering.

Fully autonomous cars are mapping out pedestrians, cars & other hazards many times a second, in all directions. If a bug is found, it will be fixed for all autonomous cars... no more teens making the same mistakes over & over again.

I can't wait for full autonomy.
 

·
Registered
Joined
·
1,443 Posts
Discussion Starter · #5 ·
I have to agree with @phr00t - humans are less reliable than machines. Too many variables with humans, hundreds more than a coded procedure. Also surely There will be overrides that you can activate if you feel your life is danger?
 

·
Registered
Joined
·
166 Posts
I have to agree with @phr00t - humans are less reliable than machines. Too many variables with humans, hundreds more than a coded procedure. Also surely There will be overrides that you can activate if you feel your life is danger?
I love the idea of Autonomous vehicles but think we're a way off yet, there are so many questions to answer:

If you override your autonomous vehicle and cause an incident should you be held liable?

If the autonomous vehicle causes an incident who is liable?

Is there a difference between an Autonomous vehicle causing an incident and a person causing one?

How do we measure and assure ourselves that autonomy is better?

Autonomous vehicles will have to co-exist with driven vehicles, pedestrians, cyclists etc. which is going to make getting accurate data on safety extremely difficult.
 

·
Registered
Joined
·
166 Posts
I have to agree with @phr00t - humans are less reliable than machines. Too many variables with humans, hundreds more than a coded procedure. Also surely There will be overrides that you can activate if you feel your life is danger?

I think the point is that it can't be coded in the conventional sense it will (and is) reliant upon Artificial Intelligence. This is very different from writing code to do something like LKAS, ACC or AEB.
 

·
Registered
Joined
·
251 Posts
I'm old enough (and old-fashioned enough) to remember when driving a car actually required some degree of awareness, concentration, skill and planning on the driver's part.

Being a staunch manual-transmission holdout (three of my four cars are manuals) in a country in which automatics are so prevalent that very few people even know how to drive a manual gearbox is a daily exercise in frustration for me. Most American drivers today could not care less what goes on under the hood, inside the transmission, at the far end of the steering column or at the points where the tire contact patches meet the road. They just want to get where they're going with minimal personal involvement, without having to think about it and definitely without having to ever take their eyes off their cell phones.

I view systems such as LDW, AEB, BSW, SCC and all the other safety tech that adorns new cars today, including some trim levels of the Ioniq, as no more than driving aids. To my mind, these pieces of hardware and software do not substitute for driving skill (although I fear that's the way they're used, mostly). Everyone makes mistakes, and these systems may keep an otherwise skilled and competent driver from getting into trouble due to a moment of inattention.

But as for sitting back in oblivious unconcern and being a passenger while these systems control the car from origin to destination? Not for me, not now and not ever. To paraphrase the National Rifle Association, "They'll take my non-autonomous car from me when they pry the steering wheel out of my cold, dead hands!"

Just my opinion. YMMV, of course.
 

·
Registered
Joined
·
162 Posts
but the code is only as good as the person who wrote the code
The code isn't written by one person at one time. It is continuously improved over thousands of miles with the input of the industry's finest.

To answer your concern another way, think of this line of code:

A = B * 35752.12 / 67;

The computer will figure that out in a nanosecond, 100% of the time. The computer is better at it than the human.

All concerns about liability are minor compared to saving lives. The insurance companies & manufacturers will have to figure that one out.

Hopefully enough people are willing to accept autonomy to really cut down on accidents. The fewer human drivers on the road, the better. The public roads are not a good place to be recklessly selfish.
 

·
Registered
Joined
·
3,473 Posts
Millions of lines of code required for autonomous driving that is constantly tweaked will always have glitches needing further updates. Never ending.
 

·
Registered
Joined
·
162 Posts
Millions of lines of code required for autonomous driving that is constantly tweaked will always have glitches needing further updates. Never ending.
True. Autonomous driving only needs to be better than human drivers, which it arguably already is. Perfection is the goal. Zero mistakes. That is why updates will never end, not because it is fundamentally flawed or "dangerous."
 

·
Registered
Joined
·
1,443 Posts
Discussion Starter · #13 ·
Autonomous vehicles will have to co-exist with driven vehicles, pedestrians, cyclists etc. which is going to make getting accurate data on safety extremely difficult.
Pedestrians are over rated. I think the world would be better off if there was less of them around. >:)

And I have to agree with @phr00t again... two major flaws to humans, ego and self importance. Probably the same two things that cause most car accidents, the other being idiocy.. whatever form that takes. Usually an idiot using their phone at the wheel. I despise that entirely. If i see someone looking down at their lap as they are driving 50mph I usually give them quite a look of disgust as I drive past. It is becoming more common and an autonomous car would take the wheel out of these idiots hands, literally, and I for one, cannot wait until that happens.

As for whether or not I would buy one, I would want the option. So if I'm on a long journey and I'm getting tired, or need to check an email, turn auto pilot on and do whatever. But then I like to be able to take back control. And I'm thinking that's not how these autonomous cars are being spec'd?
 

·
Registered
Joined
·
162 Posts
Autonomous cars will likely allow overriding for a long time. The public simply will require it during its rollout. However, I suspect over time, people will naturally stop driving themselves when they realize all the better things they could be doing with that time. To add to that, an overwhelming majority of crashes will end up at the hands of human drivers. Combine those things together, and human driving is on its way out... it will just take a long time for the public to adjust.

I also suspect a new business will emerge, where you can drive cars on closed tracks for fun. No opposing traffic, proper spacing between cars... so you can drive for fun & as selfishly as you want, without putting others at risk.
 

·
Registered
Joined
·
3,473 Posts
True. Autonomous driving only needs to be better than human drivers, which it arguably already is. Perfection is the goal. Zero mistakes. That is why updates will never end, not because it is fundamentally flawed or "dangerous."
I don't think there are any vehicles capable of true autonomous driving from city speeds (most) to highway (driver aids). Easy for them to be stuck in a logical loop as well that is simple for humans to solve (happens on simple left or right turns), or unique road conditions it has never seen. For example, how could you program a car to deal with a moving truck blocking half of a city street that will be there for hours. Two easy solutions for humans, none for a program. It will eventually be better than average drivers, but will never be any better than a good driver that is alert and has a car fitted with advanced safety features.
 

·
Registered
Joined
·
162 Posts
I don't think there are any vehicles capable of true autonomous driving from city speeds (most) to highway (driver aids). Easy for them to be stuck in a logical loop as well that is simple for humans to solve (happens on simple left or right turns), or unique road conditions it has never seen. For example, how could you program a car to deal with a moving truck blocking half of a city street that will be there for hours. Two easy solutions for humans, none for a program. It will eventually be better than average drivers, but will never be any better than a good driver that is alert and has a car fitted with advanced safety features.
Google's Waymo already has cars that go from city speeds to highway speeds. Tesla does too. It isn't easy for them to get stuck in logical loops -- if it was, it'd be easy to detect it in the millions of miles those projects have driven & get those bugs fixed.

You may have gotten the example of a blocked truck (or bus) giving the Nissan trouble at 8:50 here:


Notice how the engineer says "current limitation" & one of the reasons autonomous vehicles are still in development. If it is an easy solution for a human, the engineers just need to find what makes it easy to detect for a human & make it into the program. Saying there are "none for a program" is why you'd never become an autonomous driving engineer -- you need to be open to finding a solution before you will find one. Look for the 4 way flashers? Slowly creep around & look for a safe way around after waiting awhile? Take hints that the bus is very close to the curb?

A good human driver can not look in all directions at the same time. A computer can. A good human driver has a reaction time in the hundreds of milliseconds. A computer's reaction time is in nanoseconds. A "good" human driver is never 100% alert. A computer is.
 

·
Registered
Joined
·
134 Posts
But a computer can only respond to a situation that it has been programmed to respond to. If it comes across a situation that it hasn't been programmed to deal with then there is likely to be an accident. A driver hover can use their experience and judgement to make a decision which might prevent an accident so ultimately the driver needs to be in control and use the new safety driving aids to assist them.
 

·
Registered
Joined
·
1,443 Posts
Discussion Starter · #18 ·
It takes a lot of code, granted, but every scenario can be coded or coded with variables that allow for a wider spectrum of incidents in which it must react in a certain way. I'm still 100% behind technology. Been in the business too long to not be a believer but I also know developers are not perfect and again there is the flaw... humans. Bet if it was a robot coding.........
 

·
Registered
Joined
·
162 Posts
But a computer can only respond to a situation that it has been programmed to respond to. If it comes across a situation that it hasn't been programmed to deal with then there is likely to be an accident. A driver hover can use their experience and judgement to make a decision which might prevent an accident so ultimately the driver needs to be in control and use the new safety driving aids to assist them.
Google Waymo alone boasts: "2 million miles self-driven. Since we started at Google in 2009, we’ve accumulated the equivalent of over 300 years of human driving experience, largely on city streets. That’s on top of 1 billion simulated miles we drove just in 2016."

If a situation presents itself that hasn't been discovered over millions of miles, what makes you think a human, with all of their human limitations (blind spots, reaction time etc.) will perform any better? If something is "outside the box", an automated car still recognizes objects & their movement around the car, and will pick the safest path to avoid them all.

A human driver using their experience & judgement is exactly what I'm worried about. Humans, time & time again, make terrible split decisions. After a bad decision is made, every other human is susceptible to making the same mistake. Automated cars can fix bugs across a whole fleet to never repeat the same mistake again.

It doesn't feel right to trust a computer when you've trusted yourself behind the wheel for decades. I get that. Over time, you & the public will ease into it as the technology rolls out, and you get to see it doing its job. There will be automated accidents, sure, and they will get tons of attention. Just imagine the news if we covered every human accident like that...
 

·
Registered
Joined
·
1,443 Posts
Discussion Starter · #20 ·
A human driver using their experience & judgement is exactly what I'm worried about. Humans, time & time again, make terrible split decisions. After a bad decision is made, every other human is susceptible to making the same mistake. Automated cars can fix bugs across a whole fleet to never repeat the same mistake again.
And that's truth right there. Humans are just as predictable in the sense that we, like a computer system, have the same sensors and abilities but unlike a computer system, we react differently based on our character, mood, consciousness at the time, etc. So we are probably less reliant.

Also like phr00t says, bugs can be fixed via patch pushed out over a massively connected network. Humans can't be patched although some days I really wish they could be.
 
1 - 20 of 52 Posts
Top