In case you want to keep score on the Google driverless car accidents, they provide a monthly report that documents them.
Until the end of May 2015, Google’s self driving cars have recorded 2 million miles and been involved in 12 accidents. In 10 of them, the Google SDC was rear ended, mostly at traffic lights or stop signs. In a different part of the Google SDC website, they discuss situations when the SDC becomes “confused”. Fortunately, all accidents were low speed and no one was injured. I know the regulars have strong opinions on this subject, and I wanted to give you a chance to put your spin on it.
I realize that in a rear end collision the car in the rear is at fault, but often it is the car in front that causes the incident by driving unpredictably.
I m sure that we could come up with a safe self driving system, however it will not be practicle in the near, or distant future. we would have to remove almost all human drivers from the road and take all control from human passengers. we would need a totally revamped highway and road system, complete with a new nationwide electronic infrastructure. a whole new system for hauling freight and a whole new way of doing road construction and repair projects.
seeing as how we can t even keep the current bridges and roads in good repair I just don t see it happening except in the odd city here and there
Google says their reason for the SDCs is altruistic, but I’m too cynical to buy that. I think they want to use SDCs to replace the cars with drivers they use for street views on Google Maps. I’m sure they don’t pay much to those drivers, but even minimum wage is likely more than it would cost to run SDCs over the long term. They can’t export those jobs.
As the future slowly becomes the present, I’ll just have to learn to adapt. But until then, I’m not totally comfortable with them. It’s my belief that driverless vehicles will ultimately fill the roll of taxicabs, delivery vehicles, curriers and such in highly populated urban areas, and they’ll probably have a driving record far better than the average human in these environments.
But even if the technology were flawless, there’s still be problems to overcome… such as speed limits. In a city, where such a vehicle can orient itself to the speeds of the vehicles around it, it might work fine. But on a secondary highway, will it do the speed limit and cause a rolling road block, or will it try to find some appropriate speed limit? Most humans go faster than the speed limit. How will it adjust to that? Will it go faster and faster until it gets behind another vehicle to orient itself to? If a vehicles comes up behind it, will it speed up? Will it become confused by a vehicle passing it?
It doesn’t seem to be confused by the issues you mention, Mountainbike. It doesn’t seem to understand erratic behavior by other drivers. Two of the accidents involved another driver veering into the Sid elf the SDC and th either driver pulling the nose of his car into the path of the SDC. I well imagine that Google is working on that, but there a huge number of erratic behaviors to account for. That’s what bothers me most.
Given the dismal performance of such a big percentage of the driving public, it’s kinda hard to imagine how SDCs could possibly fail to be an overall improvement.
I wonder how the liability problem might be solved though?
Trouble is, JT, that one cannot drive ten miles without witnessing numerous incidents of erratic behavior. We humans just deal with it and go on our way, not even thinking about it, but a program needs to have the capacity to adjust to it based on its programming.
When I was teaching my daughter to drive, I used to tell he that she would never go out on he road without encountering bad decisions by other drivers, and her ability to be aware of and anticipate these would be the difference between getting in an accident and not getting in an accident. As we’d drive, I made it a point to mention it every time I saw other drivers create a potential accident situation. I even amazed myself at how commonplace they are.
Google says their reason for the SDCs is altruistic, but I'm too cynical to buy that. I think they want to use SDCs to replace the cars with drivers they use for street views on Google Maps. I'm sure they don't pay much to those drivers, but even minimum wage is likely more than it would cost to run SDCs over the long term.
It would probably take several decades to equal the cost of R&D on the google cars. Last I heard it was well over $1BILLION. I can’t see the cost of drivers even approaching that for decades.
I don’t “buy” the altruism statement. I’m guessing that Google is exploring new ways their technology can interface with cars, and this is a demonstration of the possibilities. My guess is that “Google-equipped” cars will be in showrooms soon… whatever the heck that might mean. I’m sure the revenue stream will be considerable.
And there’s another possibility… other world markets. I don’t yet know what I mean by this, but a lot of new technology is aimed at China, Europe, and cities in these places that don’t even come to mind for us.
“I wonder how the liability problem might be solved though?”
That’s a solved issue. A captain of a ship or airplane that utilizes an autopilot retains responsibility monitoring the performance of same, and taking over if/when the autopilot malfunctions. Saying “it was the autopilot’s fault” doesn’t hold water, legally.
Given that Google’s “self-driving car” is really an over-hyped autopilot for a motor vehicle…I don’t see where we’re entering uncharted legal waters, here.
As the law stands, right now, SDCs are AIDS to operating a vehicle. On whatever incident reports were filed in the Google car accidents, the driver of said vehicle was the guy sitting in the front left seat. If any “points” were assigned the accident…he’d be the guy getting 'em.
NOW, because Google has this “pie-in-the-sky” grandiose opinion re: how their autopilot “isn’t really an autopilot”…they indulge these mental musings about how to handle liability. Baloney…it will be handled as it always is handled: the Captain in charge of the vehicle accepts ultimate responsibility, regardless of whether or not he was directly manipulating the controls or not.
Heck, Capt Hazelwood was fould liable in a crash where the Third Mate was at the helm and wrecked! The precedent of “the Captain has final authority, and accepts ultimate responsibility” is an OLD point of law; been around longer that any one alive today has been on this earth. If it were to be handled any differently, NEW laws would have to written to AMEND this established point of law.
(Of course, I hope such legislation never is written, much less enacted.)
I understand, JT.
I just can’t figure their “angle”. No company spends that kind of research money on anything that isn’t consistent with their business plan, and I doubt if Google wants to get into car manufacturing, so I’m left puzzled. There has to be an “angle”. I suspect they somehow wan to link their product with automobiles, I just can’t figure how.
I am just picturing the car stopped in the middle of a major intersection, not moving because the self-drive is malfunctioning, you have to call tech support/in India; and the guy keeps going down the check list; Is the door closed? Is the key with you?..
I understand, JT.
I just can't figure their "angle". No company spends that kind of research money on anything that isn't consistent with their business plan, and I doubt if Google wants to get into car manufacturing, so I'm left puzzled.
You never heard of Digital Equipment Corp. When they were near peak in the late 80’s and early 90’s they were spending MILLIONS on projects and research they knew had no chance of ever making hem money. There were many R&D projects in HOPE that something actually panned out.
Most companies can’t do that kind of research…but the big companies with lots of cash on hand and money to burn can. IBM did it…Microsoft still does it…so does Google and Amazon and even Facebook.
Google doesn’t have to build cars to make a business out of it. They can hold patents and sell software and the knowledge of what sensors to use and how to integrate them into the SDC. The initial agreement to purchase rights to use the technology won’t cover the bills, but 20 years of royalties can go a long way towards paying for the research. As @MikeInNH said, the research branch of an organization is not always a money-maker. But they can devise products that extend the life of a business and maybe help it grow. There is no such thing as standing still in business. That is like giving up and going home. The SDC projects is well publicized, and I’m sure that Google has many other projects they are working on.