|
Oh my god, we've had GM tomatoes, GM microbes, now GM f'ing cars.
Where will it end
"If you don't fail at least 90 percent of the time, you're not aiming high enough."
Alan Kay.
|
|
|
|
|
Unfortunately, unless you want to drive a classic car anyway, all the other auto makers are drinking the same koolaid and adding LTE modems and "value added" services on top of them similar to what OnStar's been doing for the last 20 years.
Did you ever see history portrayed as an old man with a wise brow and pulseless heart, waging all things in the balance of reason?
Is not rather the genius of history like an eternal, imploring maiden, full of fire, with a burning heart and flaming soul, humanly warm and humanly beautiful?
--Zachris Topelius
Training a telescope on one’s own belly button will only reveal lint. You like that? You go right on staring at it. I prefer looking at galaxies.
-- Sarah Hoyt
|
|
|
|
|
Depends on where they have test driven the car - which country(ies), which city(ies), which particular road(s), ...
Does it include my city, my favorite (?) road there.
|
|
|
|
|
We already jump inside self driving aeroplanes so what's the problem about cars?
Most flights are 100% controlled by software and pilots are there for the edge cases or to land and take off when they are bored
IMHO, we're still pretty far from having self-driving cars broadly available, but we're getting there.
I think the main problem is not for who's inside the car but for those outside.
Think about it, would you normally cross the road as you do today but in front of a self-driving car?
Think about insurances
If the freaking car crashes, who's fault is it?
If it accidentally runs over someone, who's fault is it?
There's also the pleasure of driving...
For long distances, cruise-control already does wonders, but I like to drive.
I would still require a car with an "auto-off" mode.
There's also the traffic problem...
Self driving cars won't solve traffic issues. You won't have to drive in the queues but you'll still have them. You'll still waist time there.
All these initiatives are geek but if there were more investment in public transportation like trains and subways (that are already safe self-driving vehicles) then the quality of life of people would really increase.
So bottom line, I'm much more in favor of self-driving public transports than this kind of "imposed" geek hype about single self-driving cars that solve probably none of our big traffic problems.
Cheers!
|
|
|
|
|
AlexCode wrote: Self driving cars won't solve traffic issues. You won't have to drive in the queues but you'll still have them. You'll still waist time there.
Not necessarily. Smart cars can talk to each other, and adapt routes to maximise flow in ways that manual drivers can't.
There has been a lot of research into queuing and such like, and it comes up with surprising results like traffic jams form for no reason[^] and slowing traffic speeds up the flow of traffic: hence the variable speed limits you get on some UK motorways.
It's a complex issue - but if you have "smart" cars you can coordinate them to vastly improve road usage and reduce jams considerably, if not even eliminate them.
Bad command or file name. Bad, bad command! Sit! Stay! Staaaay...
|
|
|
|
|
That is true but experience tells me that if the traffic is better more people will take their cars instead of the public transports.
On top of the traffic jams you also have the parking issues that won't be solved even if your car is smart enough to drop you and go park itself.
So still I think that no major issue will be solved with self-driving cars.
|
|
|
|
|
I could tell it to go park itself at home, or any place within 5 miles or so. When I want it to come pick me up, it's only a smartphone message away...
|
|
|
|
|
That's a good idea but if everyone did it, it would cause even more traffic
Again, it's geek but, in my opinion, doesn't solve any of the big city traffic problems.
|
|
|
|
|
I wouldn't say "for no reason". While there isn't an obvious one to see, there are probably more cars on the road than it should be able to handle. If you take into account that you need (or want) 2 seconds of time between each car and some people react faster than others. I'd say if every car and every driver would react exactly the same way there wouldn't be a traffic jam in this case. But as humans aren't flawless only computers could provide that.
If you're in a traffic jam yourself it's usually better to keep your car rolling slowly as the cars behind you are very likely to do the same, thus bringing back some kind of flow back.
Having "smart" cars that coordinate themselves to use various routes to improve the road usage would probably also help but it isn't necessarily the thing that has the biggest impact.
|
|
|
|
|
AlexCode wrote: There's also the pleasure of driving...
I could not have said it better. If it wasn't fun, there would only need to be 2 or 3 car models at most to suite all needs.
Now if I could just afford a few more models myself
Hogan
|
|
|
|
|
AlexCode wrote: We already jump inside self driving aeroplanes I don't think that's a valid comparison. The environment is completely different: Airplanes are flying in larger distances from one another than cars are driving from one another (even when accounting the higher speed) and airplanes fly on different altitudes. Also there are no obstacles in air-traffic (a pilot wouldn't engage autopilot when flying near ground or mountains). In consequence an autopilot-software of an airplane has sufficient time to alert the pilot to take over. The only option for an autopilot-software in a car to deal with something unforeseen would be a full braking which might or might not be a good solution.
- Sebastian
|
|
|
|
|
manchanx wrote: The only option for an autopilot-software in a car to deal with something unforeseen would be a full braking which might or might not be a good solution.
Not necessarily; a good driver can also accelerate out of danger.
I agree that building driverless cars is a non-trivial task, but I also believe that in the long run - they will be safer than manual driving. That does not mean that accidents will disappear; only that they will become rarer.
If you have an important point to make, don't try to be subtle or clever. Use a pile driver. Hit the point once. Then come back and hit it again. Then hit it a third time - a tremendous whack.
--Winston Churchill
|
|
|
|
|
Daniel Pfeffer wrote: Not necessarily; a good driver can also accelerate out of danger. But what would be the point of a self-driving car if you'd be required to be ready to take over from the autopilot in a split second? I can't imagine that being more relaxing than driving myself, rather the contrary.
|
|
|
|
|
I expressed myself poorly. I meant that the software could be taught to accelarate just as easily as it could be taught to brake.
Now that I think of it, this part would actually be easier to program than crash avoidance in an aircraft - only 2 dimensions to play with...
If you have an important point to make, don't try to be subtle or clever. Use a pile driver. Hit the point once. Then come back and hit it again. Then hit it a third time - a tremendous whack.
--Winston Churchill
|
|
|
|
|
True, but it's also true that you're nowhere to be saved in case of accident or technical error.
In a car, and as everyone is stating above, self-driving will be slow and clearly under the speed limits on which most cars protect its occupants pretty well.
I fully agree that planes have less variables to take in consideration but for most people the perception is that flying is more "dangerous" than driving, even though statistics point clearly otherwise.
|
|
|
|
|
Devices are being sent to market with Beta software on the device, watch with a tiny bug is not a big deal sending a software update OK (or what we have come to expect).
Time to ship in a competitive industry will always be driven down as much as possible.
http://doubin.blogspot.com/
|
|
|
|
|
Modern cars (and aircraft) already include a lot of "drive (fly) by wire" technology, which can go horribly wrong (see Toyota's braking issues, a few years back). While bad software has the potential to cause crashes, it also has the potential to reduce them - no more drunken driving, for example.
On the whole, trying to improve software is much easier (and much less frustrating) than trying to improve people.
If you have an important point to make, don't try to be subtle or clever. Use a pile driver. Hit the point once. Then come back and hit it again. Then hit it a third time - a tremendous whack.
--Winston Churchill
|
|
|
|
|
Yes, we are going to improve the software, meanwhile they hopefully scrape our remains off that tree the software smacked us against.
The language is JavaScript. that of Mordor, which I will not utter here
This is Javascript. If you put big wheels and a racing stripe on a golf cart, it's still a f***ing golf cart.
"I don't know, extraterrestrial?"
"You mean like from space?"
"No, from Canada."
|
|
|
|
|
And how does this differ from the results of car accidents caused by human driver error?
If you have an important point to make, don't try to be subtle or clever. Use a pile driver. Hit the point once. Then come back and hit it again. Then hit it a third time - a tremendous whack.
--Winston Churchill
|
|
|
|
|
The results are the same. I just have more confidence in a driver's abilities than in a mindless machine.
At the moment I'm looking at the firmware for the control module of a model helicopter. It's open source, a real little autopilot. Last summer I lost one of the servos at the swashplate while I was flying. The result usually is a crash. My improvised reaction prevented that and then I even managed to land it. A rough landing, but one without any further damage. Most important of all, no people were in danger of being hit by a crashing heli out of control.
Now, how shall I teach the firmware of the autopilot to detect the loss of any one of the three servos and perform a controlled landing with the broken servo frozen in its last position and a totally different model of control?
As it is now, the program will try to compensate and not take into account that one servo is gone. Or I can keep on flying myself, which is a lot more fun.
The language is JavaScript. that of Mordor, which I will not utter here
This is Javascript. If you put big wheels and a racing stripe on a golf cart, it's still a f***ing golf cart.
"I don't know, extraterrestrial?"
"You mean like from space?"
"No, from Canada."
|
|
|
|
|
CDP1802 wrote: I just have more confidence in a driver's abilities than in a mindless machine.
A mindless machine has some advantages - it never gets tired, drunk, or distracted.
I agree that the current state of the art is inadequate for driverless cars. Where we appear to disagree is on whether the state of the art will ever advance sufficiently.
If you have an important point to make, don't try to be subtle or clever. Use a pile driver. Hit the point once. Then come back and hit it again. Then hit it a third time - a tremendous whack.
--Winston Churchill
|
|
|
|
|
Personally, I think I would approach the problem with some kind of evolved (GA) neural network, with the initial training (both good and bad results - you have to do both) done inside a simulator. Once the initial training is done, you could probably then continue training using a real machine with methods to fake the issues. With enough training (and constant update during regular flight), the system likely could even recover somewhat gracefully with a problem it hadn't seen before (with the exception of a complete and total failure of all systems - not even a human can recover from that).
|
|
|
|
|
I love evolutional algorithms, but they have the devilish habit tp follow your directives to the letter and still do all kinds of things you never intended.
The language is JavaScript. that of Mordor, which I will not utter here
This is Javascript. If you put big wheels and a racing stripe on a golf cart, it's still a f***ing golf cart.
"I don't know, extraterrestrial?"
"You mean like from space?"
"No, from Canada."
|
|
|
|
|
be drip-fed intead of eating?
Geek code v 3.12
GCS d--- s-/++ a- C++++ U+++ P- L- E-- W++ N++ o+ K- w+++ O? M-- V? PS+ PE- Y+ PGP t++ 5? X R++ tv-- b+ DI+++ D++ G e++>+++ h--- r++>+++ y+++*
Weapons extension: ma- k++ F+2 X
|
|
|
|
|
... it is from Apple, it probably won't let me in and I'd have to walk behind.
If it is from Microsoft, I would wait till version 3+.
|
|
|
|