Results for “driverless car” 101 found
Driverless Cars May Already Be Safer Than Human Drivers
Tim Lee runs the numbers:
Waymo and Cruise have driven a combined total of 8 million driverless miles, including more than 4 million in San Francisco since the start of 2023.
And because California law requires self-driving companies to report every significant crash, we know a lot about how they’ve performed.
For this story, I read through every crash report Waymo and Cruise filed in California this year, as well as reports each company filed about the performance of their driverless vehicles (with no safety drivers) prior to 2023. In total, the two companies reported 102 crashes involving driverless vehicles. That may sound like a lot, but they happened over roughly 6 million miles of driving. That works out to one crash for every 60,000 miles, which is about five years of driving for a typical human motorist.
These were overwhelmingly low-speed collisions that did not pose a serious safety risk. A large majority appeared to be the fault of the other driver. This was particularly true for Waymo, whose biggest driving errors included side-swiping an abandoned shopping cart and clipping a parked car’s bumper while pulling over to the curb.
Cruise’s record is not impressive as Waymo’s, but there’s still reason to think its technology is on par with—and perhaps better than—a human driver.
Human beings drive close to 100 million miles between fatal crashes, so it’s going to take hundreds of millions of driverless miles for 100 percent certainty on this question. But the evidence for better-than-human performance is starting to pile up, especially for Waymo. And so it’s important for policymakers to allow this experiment to continue. Because at scale, safer-than-human driving technology would save a lot of lives.
Driverless cars never break the speed limit, the driver is never drunk, nor distracted by their cell phone or the fight they had with their spouse. Another advantage that people might not think of is that these cars are far better for cyclists as Parker Conrad notes:
It’s so, so obvious to anyone riding a bike in SF that autonomous vehicles are WAY safer for bicyclists than human drivers. They see me every time; human drivers constantly turn right into the bike lane without thinking.
Why? Because driverless cars literally have eyes in the back of their heads.
Driverless cars are in general less good at edge cases but the advantages add up.
I would qualify this only slightly by noting that some locations are more difficult than others and while San Francisco is quite difficult terrain, Phoenix, Arizona was chosen because of flat terrain and sunny weather. Still, the bottom line is absolutely correct. Driverless cars are safer and more capable than many people think and we should always measure their defects relative to realistic alternatives and not to some idealized notion of perfection.
Truly driverless cars
California regulators have given the green light to truly driverless cars.
The state’s Department of Motor Vehicles said Monday that it was eliminating a requirement for autonomous vehicles to have a person in the driver’s seat to take over in the event of an emergency. The new rule goes into effect on April 2.
California has given 50 companies a license to test self-driving vehicles in the state. The new rules also require companies to be able to operate the vehicle remotely — a bit like a flying military drone — and communicate with law enforcement and other drivers when something goes wrong.
That is from Daisuke Wakabayashi at the NYT, via Michelle Dawson.
Will we understand why driverless cars do what they do?
A neural network can be designed to provide a measure of its own confidence in a categorization, but the complexity of the mathematical calculations involved means it’s not straightforward to take the network apart to understand how it makes its decisions. This can make unintended behavior hard to predict; and if failure does occur, it can be difficult to explain why. If a system misrecognizes an object in a photo, for instance, it may be hard (though not impossible) to know what feature of the image led to the error. Similar challenges exist with other machine learning techniques.
That is from Will Knight. This reminds me of computer chess, especially in its earlier days but still today as well. The evaluation functions are not transparent, to say the least, and they were not designed by the conscious planning of humans. (In the case of chess, it was a common tactic to let varied program options play millions of games against each other and simply see which evaluation functions won the most.) So when people debate “Will you buy the Peter Singer utilitarian driverless car?” or “Will you buy the Kant categorical imperative driverless car?”, and the like, they are not paying sufficient heed to this point. A lot of the real “action” with driverless cars will be determined by the non-transparent features of their programs.
How will regulatory systems — which typically look for some measure of verifiable ex ante safety — handle this reality? Or might this non-transparency be precisely what enables the vehicles to be put on the road, because it will be harder to object to them? What will happen when there is a call to “fix the software so this doesn’t happen any more”? To be sure, adjustments will be made.
More and more of our world is becoming this way, albeit slowly.
For the pointer I thank Michelle Dawson.
What kind of driverless cars do people want?
…the surveys also revealed a lack of enthusiasm for buying or using a driverless car programmed to avoid pedestrians at the expense of its own passengers. One question asked respondents to rate the morality of an autonomous vehicle programmed to crash and kill its own passenger to save 10 pedestrians; the rating dropped by a third when respondents considered the possibility of riding in such a car.
Similarly, people were strongly opposed to the idea of the government regulating driverless cars to ensure they would be programmed with utilitarian principles. In the survey, respondents said they were only one-third as likely to purchase a vehicle regulated this way, as opposed to an unregulated vehicle, which could presumably be programmed in any fashion.
That is from an MIT press release, here is the background:
The paper, “The social dilemma of autonomous vehicles,” is being published today in the journal Science. The authors are Jean-Francois Bonnefon of the Toulouse School of Economics; Azim Shariff, an assistant professor of psychology at the University of Oregon; and Rahwan, the AT&T Career Development Professor and an associate professor of media arts and sciences at the MIT Media Lab.
The abstract notes that if drivers are required to purchase “utilitarian-programmed” vehicles, they may be less willing to buy at all, thus postponing the adoption of what is likely to be a much safer technology.
For the pointer I thank Charles Klingman.
Driverless cars and robots are ahead of schedule
Here is the latest:
Google is sufficiently confident about its technology that its staff have discussed launching a fully autonomous taxi service in Mountain View as soon as next year, according to people familiar with the company’s thinking. The service may initially be restricted to Google employees, which might get around any legal and regulatory issues. Google has already run some tests with employees who are trained drivers.
I enjoyed this bit too:
Yet real life brings surprises no-one can anticipate. Last year, a Google car rounded a corner to find a woman in an electric wheelchair chasing a duck with a broom in the middle of the road. “We’d never tested the car against a woman and a duck,” Mr Urmson says, “and it was able to understand this was unusual, slow down, let that thing play out and then get on its way.”
Here is the Tim Bradshaw FT piece, and for the pointer I thank Michael Gibson. And Ted Craig sends me this:
General Motors Co. and Lyft Inc. will begin testing a fleet of self-driving Chevrolet Bolt electric taxis on public roads within a year, a move central to the companies’ joint efforts to challenge Silicon Valley giants in the battle to reshape the auto industry.
And here is Viv, which is supposed to be better than Siri. And here:
A robot is being designed to compete with 12th graders during the college entrance examination in 2017 and get a score qualifying it to enter first-class universities in China, according to Huaxi Metropolis Daily.
The robot will not be connected to the internet. And from the world of photography, here are robot portraits. And yet more from the FT:
US researchers have developed what they say is the world’s first surgical robot that can outperform human surgeons when operating autonomously on soft tissues such as intestines, paving the way for clinical trials.
Airbus is working with French and Japanese researchers to develop humanoid robots able to work alongside humans on its assembly lines and inside aircraft, in what would be a step change in the use of industrial robotics.
That is a lot of robot news for a day and a half.
Convexifying the driverless car
Google’s driverless car may still be a work in progress, but the potential for semiautonomous vehicles on American roads is no longer the stuff of science fiction.
By the end of the decade, a growing number of automakers aim to offer some form of hands-off-the-wheel, feet-off-the-pedals highway driving where a driver can sit back and let the car take control.
The very nature of driving, experts say, will be radically reshaped — and the biggest players in the auto industry are now vying to capture a slice of the revolutionary market they see coming within a matter of years.
From Aaron M. Kessler, there is more here.
One salvo in the fight against driverless cars
The group that stalked Anthony Levandowski, an engineer at Google X, the company’s clandestine research laboratory, calls itself the Counterforce, after a Thomas Pynchon novel. About a dozen members, all dressed in black, gathered outside the Berkeley house where Mr. Levandowski lives with his partner and two young children.
They unfurled a banner and handed out fliers detailing the engineer’s work on Google’s driverless car technology, Street View and Google Maps. The flier read: “Anthony Levandowski is building an unconscionable world of surveillance, control and automation. He is also your neighbor.”
This is still just a small and fragmented movement, as the article makes clear. I predict it will vanish, but I wouldn’t have predicted its existence in the first place.
The economy that is Dubai (a different kind of driverless car)
Thousands of the finest automobiles ever made are now being abandoned every year since Dubai’s financial meltdown, left by expatriates and locals alike who flee in a hurry because they face crippling debts. With big loans to repay to the banks (unpaid debt or even bouncing a cheque is a criminal offence in Dubai), the panicked car owners make their way to the airport at top speeds and leave their vehicles in the car park, hopping on the next flight out of there, never to return…
Ferraris, Porsches, BMWs, Mercedes are regularly abandoned at the car park of Dubai International Airport, some with loan documents and apology notes simply left on the windscreen and in some cases with the keys still in the ignition.
…Residents complain about the unsightly vehicles hogging parking spaces at the airport and sitting slumped outside their fancy yacht clubs– it’s like, so not a good look.
There is more here, hat tip goes @jscarantino. By the way, a 19-year-old in Romania may have just made driverless cars significantly cheaper.
How will driverless cars change our cities?
- Cities will greatly expand, again: Faster and more efficient transportation will convert locations that are currently too remote for most users into feasible alternatives, abundant with space. Like suburban rail in the early twentieth century and the mass consumer automobile that followed, driverless cars will generate a gradual, but dramatic expansion of cities.
- Buildings and parking will be uncoupled, freeing up valuable land: After dropping off passengers, driverless cars will independently seek parking (or their next car-share customers) and they will show up for the return ride at the tap of an app. As soon as driverless cars are common enough, the demand for adjacent parking will dwindle and parking lots in areas where land is sufficiently valuable will be ripe for conversion to other land use. As parking in high-value areas is thinned out or altogether purged, the micro-structure of cities will change – you guessed it – dramatically!
For the pointer I thank Josh Hausman.
Are driverless cars illegal?
Here is a long, 99-page article (pdf) by Bryant Walker Smith suggesting the answer might be “yes.”
My argument is less subtle than those in the footnotes of the paper. Try running a driverless car in Fairfax City, or Alabama for that matter, while sleeping in the back seat with your feet up. See what happens when you drive by an alert policeman. (By the way, if you are asleep will your driverless car respond to the police siren and pull over?)
Let’s say you sit at the wheel while the software drives, you still are pulled over, and given a ticket for “reckless driving.” You show up in court and the judge asks you what regulatory inspection or safety process your equipment has been through. I am not saying you will always lose the case or indeed always will be pulled over, but your vehicle is no longer a reliable source of hassle-free transportation, no matter what statutory arguments you may make on your behalf.
There are different notions of the word “legal,” but from a practical point of view what the police will let you get away with is surely relevant. It seems to me that your protected sphere here is quite small.
For the pointer I thank Jerry Brito.
How not to regulate driverless cars
One issue is that the laws are requiring licensed drivers to sit in the driving seat, eliminating one of the main advantages of the technology. Yet there are more problems. From Marc Scribner:
Bizarrely, Cheh’s bill also requires that autonomous vehicles operate only on alternative fuels.
And:
Another flaw in Cheh’s bill is that it would impose a special tax on drivers of autonomous vehicles. Instead of paying fuel taxes, “Owners of autonomous vehicles shall pay a vehicle-miles travelled (VMT) fee of 1.875 cents per mile.” Administrative details aside, a VMT tax would require drivers to install a recording device to be periodically audited by the government. There may be good reasons to replace fuel taxes with VMT fees, but greatly restricting the use of a potentially revolutionary new technology by singling it out for a new tax system would be a mistake.
Cheh is on the D.C. City Council.
What would it look like if we were to rewrite all of the regulations for “drivered” cars today?
Old Lady Opposition to Driverless Cars
I think driverless cars will change the design of cities, revolutionize retailing, and greatly change our driving culture, soon for example you will need a license to drive…well, you know what I mean. The scale effects on this technology are tremendous, once it works for one car it works for all. The technology won’t be expensive and it will get better every year. The technology will also get better the more driverless cars their are. Once these cars become common, for example, I expect speed limits for driverless vehicles to be substantially increased.
I do worry about lawsuits in the early years. I am not worried, however, about the following attack on driverless cars which appears to be real although it seems like something from the Onion:
One of the reasons I don’t think this will work is that the technology will be offered first as an option, like cruise control, which will appeal most to the safety conscious. The elderly in danger of losing their license, for example, may appreciate a driverless car. Personally, I would like the driverless option for night driving and I would be much happier lending my teenager the car if I could say “but only if you use the Google option!” At first when there is an accident people will ask, “did he have the driverless option on?” But soon they will start to say “if only he had the driverless option on.”
I do think, however, that technologists should change the name to the electronically chauffeured vehicle. Electronically chauffeured vehicles will appeal to the affluent, the influential and the productive.
The ubiquitous Daniel Lippman gets the hat tip.
Clifford Whinston on driverless cars
Here is one good point of many:
Driverless cars don’t need the same wide lanes, which would allow highway authorities to reconfigure roads to allow travel speeds to be raised during peak travel periods. All that is needed would be illuminated lane dividers that can increase the number of lanes available. Driverless cars could take advantage of the extra lane capacity to reduce congestion and delays.
Another design flaw is that highways have been built in terms of width and thickness to accommodate both cars and trucks. The smaller volume of trucks should be handled with one or two wide lanes with a road surface about a foot thick, to withstand trucks’ weight and axle pressure. But the much larger volume of cars—which apply much less axle pressure that damages pavement—need more and narrower lanes that are only a few inches thick.
Building highways that separate cars and trucks by directing them to lanes with the appropriate thickness would save taxpayers a bundle. It would also favor the technology of driverless cars because they would not have to distinguish between cars and trucks and to adjust speeds and positions accordingly.
The full piece is here.
Police pull over and stop Google driverless car in Washington, D.C.
Driverless car update
Getting lawmakers in the seat of a self-driving Prius has become Google’s M.O., according to Matthew Newton, editor of DriverlessCarHQ.com, a site dedicated to covering autonomous cars. “Google has been giving free rides to policymakers in California, Nevada and Florida,” Newton told Wired from his home base in Melbourne, Australia. “So it makes sense that they would do it in D.C.”
Eric Cantor, for one, was given a ride.