Google's self-driving cars have apparently had their first fender-bender — and it was a person's fault, the company says.
The car blog Jalopnik posted the above photo of one of Google's self-driving cars, which they identified by the rack on the roof that resembles a smaller version of the Street View setup. It appears to have rear-ended another Prius, to the obvious dismay of the people gathered around it.
But shortly thereafter, Google gave this statement to Business Insider: "Safety is our top priority. One of our goals is to prevent fender-benders like this one, which occurred while a person was manually driving the car."
So it wasn't the car's fault, apparently. Good news for those following the forthcoming wave of autonomous car technology.
Google has been driving around a fleet of self-driving cars for several months now, and recently lobbied the state of Nevada to pass a bill requiring new regulations for the unmanned vehicles. While safety is clearly one of the biggest concerns, cases like this also raise questions about fault. Autonomous cars are supposed to improve safety by preventing human error, but what happens when someone — a manual operator or a programmer — makes a mistake? How do you cite a car with no driver?
In this crash, at least, the apparent involvement of a live human person clears up any issues surrounding liability and fault — the driver can accept that, leaving the autonomous Prius blameless. Those questions will have to be answered before driverless cars take over our roadways.
But this incident is yet another example — as if we need one — of the human capacity for error. Hopefully when cars do take over, they'll be able to prevent these types of incidents on their own.
multiple inventions on this site can be coped together to create driverless car... tehy recently posted a link about nissan posting a temp driving vehicle, and the teh ground based gps that is accurate to within centimeters. so all you do is put a computer in, with ladar (light radar) and program it with teh rules of the road, which has already been done by some peop[le whilst i was watching discovry channel one day, made a big truck that could identify foreign objects, and knew the rules of the road to avoid, stop or go around. and you put that all into a car, adn remove the dashboard, put in a touch screen or dashboard mounted keyboard with scren to type in where you wantr to go and the car will drive itself, bing bang boom. autonomous cars that can idenftiy foreign obstructions, can follow teh road, and with the gpd/computer combo coupled with wifi links, can look ahead to see what traffic is like, wether its more economical to re-route and go a different way, and can monitor other cars around it to keep traffic fliwing smoothly adn efficiently. parking could be done by the car as well with teh ladar and cameras. no blind spots, no accidents,
and the car could monitor break wear, car function, and stuff liek it normally does and tell you when it woudl need repair, oil change, etc :)
We're on the brink of so many jumps in technology its impossible to conceive what will come in our lifetimes.
in some states the fault would have been on the vehicle that impacted the one in the front. Then what?
We all know that computers never fail, so why would there ever be a problem with driverless cars?
I cringe at the thought of the first time one of those things plows over a helpless child because it didn't "see" the poor kid.
We also know that GPS is infallible and cellular technology never goes out of range. What could ever go wrong?
Well for starters cell phone GPS is based off of tower positioning. Most forms of navagational GPS outside of cellular devices rely on satelite. This would make GPS way worse. The easiest thing would be to network the cars so they communicate the data of all the nearby cars. This would give precise road data based on other vehicles incase you were to loose temporary signal. It would also allow high speed corrections and non-slowing difficult mergers to be possible. Optimally it would eliminate traffic jams since almost all traffic trouble is based on speed variance. This tech would eliminate the speed varience eventually and everything would happen seamlessly. Eventually the cars will have to network together though.
I see people complain that these driver-less cars have the "potential" to be dangerous, indeed as DainBramage1991 says "I cringe at the thought of the first time one of those things plows over a helpless child because it didn't "see" the poor kid."
But here's the thing, how many HUMANS kill children/people by running them over because they didn't see them? Sure these have the potential to be dangerous, but they have yet to prove it in such a way as human drivers already have.
Here's the thing.
Most people are more concerned about having someone to blame (liability) than they are about the potential to save lives.
What's preferable? Autonomous cars that kill 1,000 people per year, and no one to blame?
Or people driving, killing 100's of thousands per year, but at least we know who to charge with a crime?
This is the unfortunate thing about human nature, having someone to fault is more important than saving (or improving) lives.
I have lost lovedones due to human errors in driving. Driving is a dangerous endeavor, life and death decisions are made every second. I am certain that there will eventually be some deaths resulting from autonomous cars but I think the level of safety will dramatically increase and the number of fatalities will decrease.
yes autonomous cars are the future. that one scene from minority report comes to mind. i would easily choose a computers judgement over a humans when it comes to somewhat difficult tasks like driving cross country. we are prone to mistakes a lot more when there are other people driving who are prone to the same mistakes. computers don't give a crap about speeding and driving while intoxicated. i say bring em on now!!
I'll tell you what I cringe at . . . .
According to NHTSA, . . . "33,808 people died in traffic crashes in 2009 in the United States . . . an estimated 10,839 people who died in alcohol-impaired driving crashes. Drunk driving fatalities accounted for 32% of all traffic deaths last year, that is, on average someone is killed in an alcohol-impaired driving crash about every 50 minutes in the U.S."
Guaranteed . . . driverless cars will not killed that many humans . . . ever!
I for one welcome our new driverless car overlords.
I suppose this care can drive itself, ok. But then a driver became behind the wheel. I like to know his words of what happen.. Yes, in life I stop short and not soon enough. I did not see the tail lights. I was on my cell phone. I was focus on the radio. A cigarette fell in between my legs. I was thinking and not focus. I just like to have simple words of this driver of what happen. We already have the AIRBUS planes that fly us humans mostly by computer. This scares the HELL out of me. I know if the computer dies it has no vested interest in life. I prefer a pilot as the driver. I understand in life we have sleeping, drunk or incorrect drivers; but at the least they have a vested interest in survival. A computerize car is just wrong on many different levels and not best for the individuals and the general public at large.
This systems will probably have so much redundancy to prevent accidents if something fails... Probably much safer than someone having a heart attack while driving or someone drunk driving, or falling asleep.
People are so afraid of CHANGES and the world needs lots of them!...
Damn talking snakes believers & Republicans...
(I just wanted to say that!)
Who do you blame when a self-driving car crash occurs? The designated driver, of course.
There is absolutely no reason to believe that a self driving car would not be required by law to have a human designated driver, sitting in the driver's seat at all times, that is fully responsible and liable.
Airplanes with autopilot are a perfect analogy. All commercial aircraft and most small private aircraft have autopilot to some degree. The most advanced ones can fly fully autonomous from take off to landing but they still ALL require a human pilot in command. And it is that pilot in command that is fully responsible (and liable) for everything. The autopilot is merely a tool for assistance that makes flying way more safer.
Just like cruise control today. If you are driving with cruise control turned on and crash into a car in front of you, you are responsible, not the computer system.
And so it will be that if you are in the driver's seat of a car that can drive itself, and that car crashes while in autopilot mode, then you, the designated driver are responsible. It's just that simple. Cars are not going to be driving around with no one in them for a long, long time.
Cars with more and more autopilot capabilities will only lead to safer and safer roads -- period.
And I for one can not wait until the technology hits mainstream.
As for Airbus's control system, I think back to the crash caused by the aircraft's computer overriding the pilot's manual input and flying the aircraft right into the ground. Real smart system there.
As for GPS, Cell, or any other radio based system, the problem is that they are RADIO based systems. Anyone who has spent any time studying radio wave propagation understands that it is inherently UNreliable, and therefore a bad choice for life-safety systems.
Yes, I know that humans tend to make mistakes (driver error, DUI, inattention, etc.). But rather I'd take my chances with a human driver who's built-in computer never crashes than rely on multi-system technology built by the lowest bidder that has a built-in rate of failure.
There is one very big thing over looked here autonomous cars would be temping targets for hacking.
If centrally controlled you'll be asking for trouble as such a central system would be a great target for cyber terrorism.
Cell phone networks and GPS signals are not exactly reliable neither are personal computer OSes.
The reason why aircraft can get away with one aircraft don't have to deal with obstacles and they are hundreds of meters apart vs just meters.
Plus every bit of software and hardware is tested and qualified it's not as simple and strait forward as most people think.
Second an airliner has to go in for inspection on a regular biases every x amount of hours.
Third aircraft have triple redundant fault tolerance systems this is ok for a jet that may sell for $100 Million but is not an option for a $20 thousand dollar road vehicle.
It also would be very bad news for pedestrians.
If the camera is dirty it will not see them.
Sebastian Thrun, who was the project leader of this self-driving car project, had deliberately trapped me in collateral with a criminal suspect named Gabriele Scheler in Stanford. People behind Thrun had systematically molested many years of my life without an end. Google's Eric Schmidt had threatened my life with a real murder case of Stanford student May Zhou [www.mayzhou.com] for sake of Sebastian Thrun during their fight with Stanford.
Investigat-ion from authorities after my tip confirms that it is people on Schmidt and Thrun's side who's behind May Zhou's murder case in order to threaten me and to terrorize Stanford. And the power on their side did try to plot a murder on me while I was in California. Before the case could be publicly clarified, neither Thrun nor Schmidt's name is clear in such plotted murder. So far, they dare not deny such accusations but pretend not seeing while publicly losing their faces.
In the past, Thrun's bosses had tried to get me work with Sebastian Thrun as a settlement of crimes from Thrun's side, but I never compromised a bit, because as I told the investigat-ors, that it is unfair to that innocently murdered girl May Zhou. It's unfair to myself as well, as Eric Schmidt, Sebastian Thrun and Gabriele Scheler's side did try to murder me while I was in California;
Who wants to work with a professor who's misbehaviors had caused the murder of an innocent student of their own school anyway.
For details please read COMMENTS Part in [read.bi/thrunscase]
[bit.ly/schmidttt] about Google's Eric Schmidt
[http://read.bi/schmitie] about Google's Eric Schmidt
Also attention to the photo evidence in the context [http://tysurl.com/mssgYn]
I would like to find out the specifics behind the crash. While the car was described to be at fault, they did mention a potential coding error or the driver of the other car.
As for the use of autonomous vehicles, I agree with them for long distance interstate driving. Many people "zone out" while driving long distances, I am not without blame there. For city driving however, I would like to see computer assisted driving. Say I see a potential accident I want to be able to easily override the car and swerve and brake. Or if I miss a potential accident, I would like the car to be able to stop itself.
There are plenty of potential problems with autonomous vehicles and human driven vehicles. With people, you have to be able to trust they are capable and mentally capable of driving a vehicle. I don't think a driver's license is enough fill those requirements, auto insurance proves that fact. With a computer, you have to worry about communication, programming, and information errors. Not to mention hacking problems. As a fix for GPS errors, transmitters could be placed in roads. We already have reflectors on road lines, couldn't be too hard to install transmitters, though funding might be a problem.
Ok, tempralstorm, from what existing radio services are you going to steal the hundreds of MHz of radio spectrum needed to make such a system work? That is assuming that funding could be had in the first place. The radio frequency spectrum is much busier than most people realize, and just installing a bunch of transmitters across hundreds of thousands of miles of roads would be a logistical nightmare for that and many other reasons.
As for GPS being infallible, if you were to follow your GPS from my road to another nearby road, it would happily direct you across a dirt trail that is passable only by ATVs and dirt bikes. And that's with a GOOD signal...
DainBramage, I am not talking about high power transmitters. Not something that would control the vehicles. Just a marker that a vehicle would follow. If a transmitter, or marker, only has a transmitting diameter or 2 feet, how would that interfere with high power radio signals.
And I never once said anything about GPS being infallible. And as for the example of the GPS, that isn't a problem with the GPS signal. That problem is caused by the maps preloaded on your GPS. We too often confuse GPS with the maps on a GPS device. All GPS does is located your position on the ground. Its up to the software on the GPS receiver to provide you with maps and a route to follow.
A hacker in this case would be the same as a guy intentionally running over people.
The hacker would be in serious effing trouble!
Sure, driverless cars are dangerous, especially if that and that, and that, and that, and that, and that, and that... Oh yeah, almost forgot that too!
BUT STILL SAFER THAN DRIVERS, and that's the whole point of it!
In the end, it couldn't be as bad as current statistics.
Computers can also be human-caused....
I don't care how many benefits this system brings, and I won't argue the benefits. I hope I'm never forced to give up my privilege of driving MY cars. I bought them, I insure them, I register them, I perform my maintenance etc etc.
tempralstorm, your last comment just proves how little you actually know about RFI. Just a hundred transmitters putting out 250 milliwatts each over a only few square miles will effectively blank out any and all worldwide radio communications on multiple bands in that region. A nationwide system would effectively make long-distance radio communications impossible.
But if polluting the airwaves so that you can nap instead of driving is what you want, that's what you would get. The RF spectrum essentially destroyed for your convenience.
I wonder what the first insurance policy for an autonomous car would look like and cost
and dainbramage or whatever did you read this?