Please register to participate in our discussions with 2 million other members - it's free and quick! Some forums can only be seen by registered members. After you create your account, you'll be able to customize options and access all our 15,000 new posts/day with fewer ads.
I don't agree with your premises at all. I do not see or agree that there are things the computer cannot know or sense. It is just a matter of engineering the right tools to do the job. Those sensors that are not available today - will be tomorrow. I just think many people have very limited vision for what can be done.
The limits of the computer are it's experience level. And because there will most likely be a database of "situations" built and shared amongst the millions of drivers, that database will expand fast. If a human can do it, a computer can do it faster and probably better than most drivers - and this does not even take in consideration the advent of AI. IMO, there is nothing a computer won't ultimately be able to do.
Computers already control or directly aid in the most difficult and advanced problems in the world - driving is not that advanced a problem - and it will be less so when human drivers who do stupid things daily are removed from the equation.
That remains to be seen if a computer can do it better than a human. In an above linked article Tesla said that California needs to improve lane markers, to stop their autopilot cars from killing people. The lane markers are good enough for millions of human drivers, but they are not good enough for the computer. That is an indication that the computer is not doing as well as a human driver.
As for that database expanding fast, it's not fast enough. There are 4 million miles of roads in the US, and half of the roads are not even passable for driverless cars, at this time, because nobody as of yet has been able to program a driverless car to operate on a road without lane markers. So they can't even start to build a database of those roads. And now they are telling us that the lane markers we have are not good enough for their self driving cars.
That remains to be seen if a computer can do it better than a human. In an above linked article Tesla said that California needs to improve lane markers, to stop their autopilot cars from killing people. The lane markers are good enough for millions of human drivers, but they are not good enough for the computer. That is an indication that the computer is not doing as well as a human driver.
As for that database expanding fast, it's not fast enough. There are 4 million miles of roads in the US, and half of the roads are not even passable for driverless cars, at this time, because nobody as of yet has been able to program a driverless car to operate on a road without lane markers. So they can't even start to build a database of those roads. And now they are telling us that the lane markers we have are not good enough for their self driving cars.
Here is a Tesla operating on a markerless road. As I said, it is just a matter of time before it is perfected.
Who would have thought it would take just a short time for Google to have mapped most of the roads in the country, I would have thought it would take decades.
" Tesla said that California needs to improve lane markers, to stop their autopilot cars from killing people. The lane markers are good enough for millions of human drivers, but they are not good enough for the computer. That is an indication that the computer is not doing as well as a human driver".
If this is a problem, what the h**l is going to happen when weather conditions get really bad. Or the sensors and cameras get fogged up, snow or mud covered. Let them have their own lanes and let them crash into one another. Just because a technology is possible, doesn't make it necessary. Unless money can be made by somebody. And let people die as they get the bugs out.
" Tesla said that California needs to improve lane markers, to stop their autopilot cars from killing people. The lane markers are good enough for millions of human drivers, but they are not good enough for the computer. That is an indication that the computer is not doing as well as a human driver".
If this is a problem, what the h**l is going to happen when weather conditions get really bad. Or the sensors and cameras get fogged up, snow or mud covered. Let them have their own lanes and let them crash into one another. Just because a technology is possible, doesn't make it necessary. Unless money can be made by somebody. And let people die as they get the bugs out.
Shhh! Every time I point out that nearly all AV development has been done in very modern city areas, industrial parks and test areas with exceptionally good, well-marked roads, someone calls me a naysayer or a Luddite.
Who would have thought it would take just a short time for Google to have mapped most of the roads in the country, I would have thought it would take decades.
That is a bug, and misuse of the software. Current software will not allow Autopilot to engage on roads without lane markers. So that is not supported by Tesla. Though apparently they are happy to let drivers do that at their own risk.
As for Google. 11 years ago Google started their Street View program with the stated goal of imaging every single road in the US. At this time there are still massive sections of the country that have never been driven by a Street View car. Worldwide Google has managed to Street View about 40% of the paved roads in the world. At that rate in another 15 or 20 years, they could have every paved road in the world included in Street View. But that would still not include the majority of roads, which are not paved. Also by the time they finish, most of the images are going to be very dated.
So far Google has only been mapping very small areas for use by their self-driving cars.
My point being that I don't think you understand how time consuming it is to map all the roads in the country. It is not a simple task. I don't think it is even practical. I have no doubt that it can be done, but keeping the maps up to date will be another matter. And if they miss even one single obstacle, it could result in someone's death.
AVs without detailed maps are being developed. Why do anti-AV people refuse to acknowledge that today's technology does not define an AV development program?
" Tesla said that California needs to improve lane markers, to stop their autopilot cars from killing people. The lane markers are good enough for millions of human drivers, but they are not good enough for the computer. That is an indication that the computer is not doing as well as a human driver".
If this is a problem, what the h**l is going to happen when weather conditions get really bad. Or the sensors and cameras get fogged up, snow or mud covered. Let them have their own lanes and let them crash into one another. Just because a technology is possible, doesn't make it necessary. Unless money can be made by somebody. And let people die as they get the bugs out.
Yep, you can scratch the snowbelt off the list for places where self-driving cars will be allowed. When/if self driving cars ever hit the market, they will be restricted to California, Arizona, Florida, and maybe then expanded to some other sunbelt areas. They will be a small niche product for certain parts of the country.
That is a bug, and misuse of the software. Current software will not allow Autopilot to engage on roads without lane markers. So that is not supported by Tesla. Though apparently they are happy to let drivers do that at their own risk.
As for Google. 11 years ago Google started their Street View program with the stated goal of imaging every single road in the US. At this time there are still massive sections of the country that have never been driven by a Street View car. Worldwide Google has managed to Street View about 40% of the paved roads in the world. At that rate in another 15 or 20 years, they could have every paved road in the world included in Street View. But that would still not include the majority of roads, which are not paved. Also by the time they finish, most of the images are going to be very dated.
So far Google has only been mapping very small areas for use by their self-driving cars.
My point being that I don't think you understand how time consuming it is to map all the roads in the country. It is not a simple task. I don't think it is even practical. I have no doubt that it can be done, but keeping the maps up to date will be another matter. And if they miss even one single obstacle, it could result in someone's death.
Google was doing this a year ago in my area, they were mapping specifically for the transition to self drive.
Its GOING To happen, too much money and resources been sunk into it over the years for them to just give up now. Trucking companies have alot of money invested in this and will darn sure see it thru.
Google was doing this a year ago in my area, they were mapping specifically for the transition to self drive.
Its GOING To happen, too much money and resources been sunk into it over the years for them to just give up now. Trucking companies have alot of money invested in this and will darn sure see it thru.
Do you have a link to trucking investing money in autonomous vehicles? I can find no source to support that.
Just because a lot of money is spent, doesn't mean it going became reality. Google spend almost a billion dollars to develop Google Glass. By now we were all supposed to be walking around with Google Glasses on our heads, but that didn't happen, and Google cut their losses.
AVs without detailed maps are being developed. Why do anti-AV people refuse to acknowledge that today's technology does not define an AV development program?
The important part of that article.
Quote:
For all the progress that has been made with self-driving cars, their navigation skills still pale in comparison to humans’.
Please register to post and access all features of our very popular forum. It is free and quick. Over $68,000 in prizes has already been given out to active posters on our forum. Additional giveaways are planned.
Detailed information about all U.S. cities, counties, and zip codes on our site: City-data.com.