Hacker News new | comments | ask | show | jobs | submit login
Google's Autonomous cars complete 300,000 miles without an accident (techcrunch.com)
288 points by tbenst on Aug 7, 2012 | hide | past | web | favorite | 263 comments

Too bad it doesn't mention if this is better than a human or not. So, I decided to figure it out...

2.9 trillion miles driven in 2009 http://www.fhwa.dot.gov/ohim/tvtw/09dectvt/09dectvt.pdf

10.8 million traffic accidents in 2009 http://www.census.gov/compendia/statab/2012/tables/12s1103.p...

If my math is correct the average american can expect to get in an accident every ~250,000 miles driven, (NYC to LA 90 times) so this does seem to be an improvement if all else is equal (which I'm sure is not the case!).

What I’m interested in is whether they counted hypothetical accidents. As far as I know, there always has to be someone in the driver’s seat, ready to step in at any time. How often did that person have to step in? Now that would be an interesting metric, especially when plotted over time.

As is, it should be pretty unlikely for the autonomous cars to cause accidents, even if they were much less reliable than human drivers (since there is always a human ready to step in).

I think autonomous cars are awesome and very clearly the future, but I find it hard to believe that humans never had to step in, especially not in the beginning. I mean, that’s why they do those extensive test drives, right? To find bugs that they can squash.

Computer driven cars also have the option of just stopping when things aren't looking good (analogous to a panic). For all we know they could stop when a leaf blows across the road - unlikely to cause an accident because of the reaction time.

The "just stop" approach isn't available for planes or nuclear power plants for example.

The "just stop" approach is also problematic on a busy highway.

I don't know where you live, but where I live the highways have lots of stopping during busy times. And of course drivers always have to be aware of stopped cars as they do happen. I suspect highways are where the automated cars have the easiest time, and it is far more likely side roads and suburban areas that are more confusing. Abrupt stops are more likely there (eg kids/pets running across the road). In the US I do believe that running into the back of another car under almost all circumstances makes you at fault.

Where I live a panic stop in heavy traffic on anything called a "highway" could get you killed. I imagine a lot of the stops we see on the highway that do not result in accidents are made safer by the network effects of lots of cars doing the same thing over time and communicating visually. It's different in nature than somebody (or some machine) getting confused and stopping suddenly in the middle of the road.

Don't get me wrong: things aren't so grim for the "just stop" approach of dealing with a problem. There's no reason an automated vehicle won't be able to communicate with other cars, warn everyone of an emergency stop (at a minimum), even find the best path off the road, depending on how serious the problem is.

Not if all of the cars can do it just as easily.

I love the mental image of a whole highway full of cars slamming on their brakes to let a leaf blow by.

How would a full switch to automated cars work anyway? Would they require different roads?

If their functionality included stopping on a dime at a duck on the road, and human reaction speed isn't equivalent, it isn't safe to human drivers. But cost and manufacturing has to factor in before we can just replace every car on the road. People still drive 20 year old cars.

I expect it will be a mixture of carrot and stick like any good phased replacement. Dedicated lanes (since they can drive 2ft from the car in front) and higher speeds because they are able to cut the reaction time down so much lower.

Eventually making it mandatory for some classes of roads (motorways / highways).

At some point in the very distant future your classic car will only be good for showing off on track days or classic car meets.

Actually, it kind of is available for planes (at least the autopilot systems). See the AF447 accident and how the autopilot disconnected itself when it started getting conflicting data from the pitot.

Agree. Did humans step-in when they noted conditions which they believed the car unable to handle and perhaps prevent the majority of likely accidents?

My suspicion is that the figures quoted around these Google cars and their safety record, if you take them at face value, lead to a ludicrous level of optimism.

I think that Google have simply done a better job of creating a self-driving car than you are willing to let yourself believe.

They have done this by attacking the problem from the 'Big Data' mindset which Google are intimately familiar with. Instead of trying to crack hard AI problems (computer vision being the most obvious) they are simply counting on being able to record and use enough information both before and during the actual driving that they simply avoid these problems.

From what I have seen this approach seems to be working. My real concern is it will not be economically viable to integrate the kind of sensor arrays they are depending on into a real car (2 radars and the laser scanner being the killers).

Indeed. The whole self-driven car isn't a new idea - various individuals and university disciplines have been trying the idea for the last decade over in Europe - it's legal to test these ideas on the road in some countries over there.

If you make the assumption that all cars will be AI, then traffic-handling becomes very, very simple. It's not obstacles or accidents that cause serious problems for traffic, it's the fact that you can't know what all other drivers are thinking.

If all vehicles behaved to a certain rule set then all vehicles actions could be predicted, and as a result designing logic to accommodate that is made much easier.

That might work if the road system were entirely isolated from the rest of the world, but it's not. There are walkers, cyclists, animals, children and other random obstacles that might exist, none of which will ever be controlled by computers.

The idea of all cars following the same set of rules even if they are all AI driven is actually way too complicated. Bugs and misunderstandings of the standard would cause deviations that the other cars would need to react to anyway.

> If all vehicles behaved to a certain rule set then all vehicles actions could be predicted.

... or, potentially, hacked.

They are already putting radars into higher end luxury cars as part of their cruise control systems. Granted, google cars involve a good bit more than those systems, but I am fairly confident that the usual trend of features trickling down into normal consumer cars will continue.

I don't think that Google would say otherwise. This is just a status report in the beginning of the program.

really? 300,000 miles without an accident is not particularly amazing for a human. I mean, it's good, but not amazing.

I work for an insurance company that specializes on auto insurance. It'd be filthy rich if people only had an accident every 300,000 miles.

The company is pretty well off, but pays 95% of the premiums in claims payment.

Doing a very back-of-the-envelope calculation, 22.5% of our insured cars had a claims payment this year (august to august), so an average insured person has a traffic accident every 4 years or so, and does way less than 300,000 miles every 4 years (more like 100,000 being generous).

I was saying not that it wasn't better than average but that it wasn't amazingly good. I'm pretty sure my mom has gone 300,000 miles between accidents; (growing up, we went on a lot of road trips.)

all I was trying to say is that I think the best human drivers can pull off three hundred thousand miles; it's not inhuman. and that's what we are going to need for self-driving cars to become widely accepted; something that would seem impossibly good to a human. And I bet they will do it, eventually.

I was responding to

>My suspicion is that the figures quoted around these Google cars and their safety record, if you take them at face value, lead to a ludicrous level of optimism.

and I don't think 300,000 miles without an accident is a ludicrous number; it's something a human can achieve, (even if it's something most humans don't.) It's a milestone, sure, but like I said, for self-driving cars to be accepted, they can't just be better than the average driver; they need to be better than the best drivers.

Ah, ok :) , I hadn't understood your statement correctly.

I'm looking forward to self driving cars because of their convenience ! Imagine being able to sleep during road trips, using them as a taxicab for places where there's no parking (and ask it to park itself :) ), sending them to pick up somebody... the possibilities are endless !!

What type of accidents though? Do people get in a number of fender benders over their lives? or head on at 65mph?

What sort of accidents will GoogleCar be prone to? Fender benders, or a head on at 65mph every 500K?

I'm very hopeful about this. Most people don't want to drive most of the time. As long as I can still drive my V12 Aston martin manually when I want to, this is awesome.

I imagine, once automated cars are safe and popular enough, they will mandate that all new cars must have it and set a date after which only automated cars are allowed on the road. It will become an offence to take over and drive manually unless the circumstance demands it. You'll be fine driving your Aston Martin, but your kids, or their kids might not be.

I expect that as the automated cars get sophisticated enough that they no longer act individually, but rather each car is in constant communication with the cars in the area so that they can coordinate maneuvers and engage in global planning, a semi-manual option might be added to the system so people can drive for fun.

They way that would work is that when you car is in semi-manual mode, it tells the others cars that a human is controlling steering and speed. The other cars will then give the human extra room (e.g., no drafting or close formation driving on the freeway, no crossing at right angles in an intersection at full speed just inches apart, or that kind of stuff that the cars will do when computers are driving all of the cars in an area).

The automated system will still be monitoring the human and can take over if the human does something bad, so the other drivers don't have to worry that a human driven car will suddenly cross the line into their lane and kill them.

This could even be turned into a game, where the automated system can score the human driver based on how well the human drove and how often the system had to intervene to keep the human from crashing.

Note that racing could also be allowed in such a system. Want to take that Aston Martin out for a race against your neighbor's Ferrari? Let the system know, and it can give you a course through the city, clear a bubble around you two, and let you go at it.

Most of these things will not be possible merely because they give you an unfair advantage. What you call "racing" I call "being at my destination much sooner", and, yes please, I would like every other car to step aside so I can get home faster. Which, of course, only means that everybody will try to abuse this.

Make "race mode" cost money then. Cheap enough that an occasional race is not ridiculous, but expensive enough that it is not practical for a commute. It probably wouldn't be an incredible advantage anyway. Your car could drive much faster if it knew exactly what every car around it was planning on doing.

At least in urban areas, you're eventually going to end up having vehicle performance arbitrated via radio by a computer network outside the vehicles. (I'm pretty sure it's much easier than trying to do everything "peer to peer," if you will) This gets you intersections where nobody has to fully stop and vehicles merely coordinate and modulate their speed, which is a pretty powerful efficiency advantage in itself.

With a system like that you could easily design a market where vehicles can choose to pay a toll to get a favorable path through traffic. You might see a few options and incentives given for choosing a more fuel-efficient path, etc.

Honestly, I'd expect semi-fun mode to be implemented to prevent human drivers from passing, thus killing the "fun" and reinforcing the "driving is tedious" meme in order to serve a greater safety goal.

I don't buy it. In that scenario, how does anybody learn to drive in the first place? Someone who hasn't spent years of their life driving manually won't have developed the skills to intervene properly in an emergency. At best, driving would, become a specialty career/hobby, like piloting aircraft.

By the time all personal transportation is automatic, it won't just be the status quo with "robot chauffeurs". The whole transportation ecosystem will have to change substantially. Self-driving cars are just a first step to get people used to (or even enamored with) the idea of automatic transportation on a large scale.

I love driving. I find it incredibly enjoyable. I don't think automatically driven cars will change that because I'd happily give up 80 - 90% of my real world driving to a computer. Driving on a congested motorway (highway) or in a busy town just isn't fun, no matter how nice the car you are driving.

I think horse riding might be the best example of the way things will go if and when automatically driven cars become a reality. Once a life skill required by everyone, now a hobby in all but the most extreme of communities.

We're already heading down that route [1]. Although I can't see it ever being an offense to manually drive your car, the car will just automatically take over the driving duties if it doesn't like the way you are driving.

[1] http://www.autoweek.com/article/20120803/carnews/120809942

You're not factoring in the Media, which will raise a big fuss about taking away their rights and robots not being safe... you get the idea. I can't imagine how people will react once there's an actual possibility a car could be driving on the road with no human controlling it.

I believe that 10.8M accident figure is an estimate counting all kinds of crashes, even the ones not reported to police.

6M police-reported crashes in 2007. 1.71 million police-reported crashes with an injury. 37,000 traffic fatalities in 2007, 30,862 in 2009.

Traffic fatality data is here with only some high-level info about non-fatality crashes: http://www-fars.nhtsa.dot.gov/Main/index.aspx

A lot of those accidents happen in rain and snow, which the Google folks admit are not as well tested. Under clear conditions on open roads, humans probably do better than 300K miles too.

That doesn't seem right to me. A person driving 10K a year will hit 300,000 miles over the course of 30 years. Does the average person really have less than 1 accident every 30 years? Especially when they are younger?

Per mile is a pretty poor measure. My dad drives at least 40K/year (long commute) and has averaged one accident per 20 years - but obviously for a regular driver following a familiar route the chances of an accident are pretty low. I would expect someone who drives less would actually have more accidents/year.

Mentally reviewing the people I'm close to, the figure seems to be off by at least an order of magnitude.

I think it includes truckers who are far safer than normal drivers on average and do a lot of driving.

I put 40k on my first car and hit stuff enough to warrant replacing something - often just a fender etc - at least 5 times. Only one fender bender when I rear ended someone at ~15 mph talking on my phone like an asshole.

I know I'm a worse driver than most. But I also wouldn't say that my similarly young friends average anywhere close to 1 such error per 250k miles. Definitely closer 25k.

I don't know much about this google project. Can autonomous cars (i) drive on the highway (ii) drive at night (iii) drive in snow or rain?

Autonomous cars have driven on the highway in real-traffic conditions [1], [2], although they had a driver at the ready in the event of any problem. I don't know about night or weather conditions for these instances.

[1] - http://en.wikipedia.org/wiki/VaMP (their citations pages aren't responding at the moment)

[2] - http://www.argo.ce.unipr.it/argo/english/index.html

They talk about weather issues in this video at 3:40.


In summary: They can handle darkness and rain, but not snow.

If you're interested, Sebastian Thrun has an intro to the topic as this course (you can preview all videos and problems)


Driving at night and the rain will probably be where autonomous cars really outshine humans. Lidar systems don't care if you are wearing all black during a night at new moon.

Snow is much harder for the time being. A snow drift falling off a car can make a google car think a cooler has appeared in the road.

Could it handle Lombard Street?

How would it react to a roadway that's covered in several feet of water due to flooding (and poor road design)? Or sinkholes? One-way streets? Freeways which change directions at different times of day? Dirt roads? (Can we make it autonomously offroad?) How will it react during an earthquake or tornado? How will it react to a bicyclist running a red stoplight and into the car's path, when the car has a green light? When the car comes across kids playing football/street hockey/baseball in the street, will it stop for them? How far out of the way will the kids have to move before the car thinks it's safe to go past?

Will these cars have systems to react when they take damage?

Google's algorithm (which Strum seems to be pollinating everywhere -- and good, because the old algorithms were failing DARPA tests) is to pre-map the road. The car knows exactly where it is and what is supposed to be there, which it also compared with what it sees now.

How will it react to a bicyclist running a red stoplight and into the car's path, when the car has a green light?

It will stop. In fact, the "failure mode" of self-driving cars will be how pedestrians can bully them by stepping in front of them so they stop.

When the car comes across kids playing football/street hockey/baseball in the street, will it stop for them?

There is already video of this. The radar/lidar isn't good enough to see a baseball, but it's definitely good enough to see a kid.

Will these cars have systems to react when they take damage?

It will probably pull over and stop if something is wrong.

Interesting case here for carjackers. You're Google car probably won't floor it in the middle of the night through a red light when an armed gang approaches your car. Then again I guess that's what "manual" is for.

Carjacking a car loaded with recording equipment that probably phones home with GPS and other environmental data probably isn't the best idea.

carjackers aren't known for their brilliance

Yes, a Google autonomous car drove down Lombard Street: http://googleblog.blogspot.com/2010/10/what-were-driving-at....

Here you can see video of self-driving cars handling Lombard Street, as well as dirt roads, off-road driving, pedestrians and bicyclists and other obstacles in the road, plus tons of information about how they work:



https://www.ai-class.com/course/video/videolecture/210 (Lombard Street at about 4:40)

One thing that strikes me is how much more these robots can potentially see than a human driver. Imagine how much better you could drive if you could see in all directions at once with multiple range-finding sensors and cameras, track dozens of moving objects simultaneously, never get sleepy or look away from the road at the wrong time, react to events in milliseconds, and learn from hundreds of thousands of miles of training data as your driving "experience".

Everything you've just said would have come up in numerous meetings before they even started planning how the system would work.

I guess another good question is: "how well do humans handle these things?"

I think the answer is: "not very well."

The original DARPA Grand Challenge was set in the desert on dirt roads. It's not that difficult to identify the road given some assumptions about how roads work (e.g., it's likely to be a slightly different color or texture)

I thought they were limited to driving on the highway in the empty parts of nevada?

I drove behind one yesterday as it returned to Google HQ in Mountain View. It was pretty exciting. It was also rush hour, and people didn't like that it tried so hard to give a safe amount of space in front and behind itself - it ended up being a little bubble of lane-changing space that moved slower than the rest of traffic.

Not that I hold any of this against the car - it was clearly doing the intelligent thing. But people... we are animals.

Weird. Your post didn't have a reply link. I had to click the permalink for it.

Anyway, I wanted to say that I read an article about a guy in Seattle that always kept a couple car lengths ahead of him. I seem to remember the article saying he was able to get home quicker that way than by behaving like the rest of the cars on the road.

Meta: there's a delay between when a comment is posted, and when the reply link becomes active. That delay is larger for deeper comment threads. I've never heard an official explanation, but one can imagine a few reasonable ones, e.g. to encourage people to stop and think in heated back-and-forth exchanges.

I too live in the Seattle area, and I've heard of this technique. Sadly, when implementing it, all you see is the angry drivers behind you who rush to pass and gobble up the traffic bubble you've been saving.

I thought that this was covered by the Traffic Waves guy who said his method deals with that problem quite well?

Yep: it doesn't happen, FAQ: http://trafficwaves.org/tfaq.html#five

- if truckers can maintain a bubble, how can it be impossible?

- yes, aggressively weaving drivers sometimes jump in ...then they jump out again.

- at merge zones, the entire point is to let people into your lane.

- In theory, everyone else wants your precious bubble. In practice, not.

...and in a burning building, we all know the best way to escape, eh? EVERYONE AGGRESSIVELY PUSH FORWARD, DON'T LET ANYONE ELSE GET PAST, COOPERATION AND TEAMWORK ARE FOR LOSERS! Yep, that certainly won't trigger a huge clogged mess where everyone dies.

That's the one. Thanks!

Seems like the safe distance in front of a robot car would be significantly less than for that of a car driven by a human. Of course, most humans I know think that the safe distance is 2 feet, which is probably less still.

They probably have to assume the human co-pilot may still need to assume control and avert disaster, and give them enough extra time to do that (beyond the time they'd need after starting out in control). With no need for failover, the robot should reliably react faster, though it's probably only reacting to immediate events rather than predicting drivers' (mis)behavior.

Most people stay way too close to the car in front, though. You're doing well as long as nothing unexpected happens. But if someone for some reason has to hit the brake really fast, you'll get a completely unnecessary pileup.

The Google cars probably also have additional margins of safety to allow for the driver to react to unexpected computer behavior _and_ actually step in and perform the maneuver himself.

  > Most people stay way too close to the car in front, though.
Yes, and people who don't care about this risk are very annoying drivers for the people who do.

I usually lightly tap the brake pedal, to let the lights come on, when they're too close behind me.

I've seen the Google self-driving car navigating within morning traffic in San Jose. There were two men with laptops in the driver and passenger seats, presumably monitoring the car and available to step in if something went wrong. It was also kind of surprising to see the roof-mounted device it used to map its surrounding. It was fairly large and rapidly rotating.

Did it look like this: http://velodynelidar.com/lidar/hdlproducts/hdl64e.aspx

Very impressive spec - as you'd expect for a $75,000 product.

The device on top is a laser scanner creating a cloud point map for spatial awareness.

that is for a car with no human in it, afaik. the google cars drive on normal roads with other cars present

To date, all Google autonomous cars have had two occupants - a driver (who is strictly hands-off - she's only there in case something goes wrong) and a tech who monitors the data to make sure things are proceeding as expected.

Only problem is sample size really. We need 2.9 trillion miles on Google cars with x accidents then.

If the Google Car had gotten into 100 accidents over 300 miles, we easily could claim it much worse than most drivers. However, as with most rare events...

I think the true rate of human accidents is higher than 1 per 250K miles - many are simply never reported to the police or DMV, especially because many states have a cost threshold ($750 in California, higher in other states) or require an injury to be reported

I'm guessing that more accidents happen in intersections than on the highway. My understanding is that Google's self-driving cars are currently limited to highways and known-course driving.

I don't think it's going to do a very good job driving in San Francisco city streets (but then, who does?)

From what I know it's always known-course. If you were to get a google car today for your commute (you can't), you would first drive to and from work yourself, probably more than once, so it can map everything.

Knowing where there should be a stop sign and traffic lights significantly reduces risks. The computer still needs to be able to learn when new ones appear, of course.

It would also help with some roads that are incorrectly marked on the GPS. There is a highway in ND that has a 4 mile stretch that has a miss marked road. Multiple GPS maps show the road about 100 yards west of where it actually is. Damn annoying driving 4 miles with the GPS constantly saying "Get back on the road". Thought it was just a garbage GPS map, but it happened on a couple of models.

Guess I really don't have to worry until they figure out how to get the car to drive on snow covered roads.

That's 10.8M crashes. The vast majority aren't really "accidents", since they are caused by human actions.

Census needs to fix their reports to be consistent with Department of Transportation terminology...

The most amazing part is that autonomous cars have completed 300,000 miles at all. Congratulations to the engineering team!

Around fifteen years ago, I read as a child with fascination an account in Scientific American about an automated highway project where sensors were to be placed every few feet on the road, and cars would follow them, until a driver retook control at the end. I imagined a future network of roads in which cars read from sensors to determine where they were, and was kind of saddened that governments didn't immediately start putting these sensors into highways. This is such a cooler solution, one that doesn't depend on a parallel development of infrastructure, one that would presumably take lots of bureaucratic steps that are naturally associated with usage decisions on publicly owned land (which makes sense).

What this reminds me of, rather spectacularly, is that if one method of getting to your solution fails, either because those sensors didn't work out, or the bureaucracy didn't, if you come up with a solution, and say "oh by the way I did it already", it's a lot harder for anyone to ignore it.

> This is such a cooler solution, one that doesn't depend on a parallel development of infrastructure, one that would presumably take lots of bureaucratic steps that are naturally associated with usage decisions on publicly owned land (which makes sense).

And taxes. Google isn't being funded by our dollars, except through a series of indirections.

I wonder if they get a tax credit for this kind of research?

Seems possible, and certain kinds of tax credit may as well be subsidies. All speculation on my part, though. I wonder if anyone knows?

A tax credit for a research project which until very recently was known only to a small group of people in Google's super-secret "project X" labs?

So secret it sprang from the government-funded and highly-publicized DARPA challenge?

I think trains/metro rails are just outright much more cost effective, given it's not personal transport.

Actually, not when you factor in the need to serve unpopular routes and non-Busyntimes. That's a lot of trains and buses running nearly empty.

Small, on-demand personalised mass transit has the potential to be the most efficient transport system.

That's really just a negative externality of inefficient suburban development. If the 'burbs didn't exist in the first place, there would be no need to subsidize their barely-used mass transit services.

The Farebox recovery ratio even for big European cities is well below recovery rates for mass transport.

Only a few Asian cities get more than their public transport systems cost to run back. This page has more info:



"That's really just a negative externality of inefficient suburban development." Shouldn't public transit cover rural areas and small towns also? Or is the only way to get an efficient transport system to have the entire world living in huge urban centres? Personal mass transit could allow a much wider range of city sizes to be served with relative efficiency.

Mass transit doesn't need "huge urban centers" and indeed can and does serve cities of all sizes. What it requires, though, is a minimum density. You can't serve a small, sprawling city. You can easily serve small, compact cities. Compare, for example, the transit effectiveness between Zurich and Tulsa, cities of comparable population.

Especially if you can finally get high-occupancy vehicles right. Something closer to a 6-passenger vehicle with bus-style seating... if we're very lucky, only with people at most one indirect hop away in your social network?

Or smaller vehicles driving inches apart to get into the slip stream, I imagine that has the potential to be quite efficient (and flexible) too

What? No way, trains/metro are an awful system that needs to be displaced by robot cars as soon as possible. A train is a large vehicle taking a lot of people to somewhere that no one wants to go. A robot car can pick me up at my door and drop me off at the front door of my job. It would only need to drive when there are people who want to go somewhere instead of "every 15 minutes regardless of how few people use it".

It depends on the population density, on whether you have to (and can) build it underground, etc.

Autonomous cars are the next big tech that changes the world - I put them up there with computer, Internet, and smart phone.

When economists think about technology, they look at it as a productivity booster. From that angle, robot cars will have a huge impact - so many commute hours will be transformed to work, or even leisure, which also boosts productivity.

So many smart people are about to be given a 10% or better bump in time to make and to do.

Eating food from a tube could also give us a lot more free time but we don't do it because it is not an enjoyable experience. We not only enjoy the flavors and textures that regular food provides, we also enjoy the rituals that accompany eating. In this regard driving cars is somewhat similar. A lot of people enjoy driving and many actually find it therapeutic. Some exceptions do exist, like driving in heavy traffic, driving long distances, and driving to places where parking spaces are limited, but better public transportation would solve those problems much more efficiently than autonomous cars ever will.

So, while autonomous cars could and probably will replace regular cars in places that have regular roads and good infrastructure they will definitely face a lot of resistance from consumers.

Not all eating is for enjoyment, though, just like not all driving is for enjoyment. People eat fast food which is certainly not as enjoyable as a well-prepared meal because of the convenience. People may like to drive sometimes (nice day, low traffic, they're on vacation - whatever), but it would take an unusual dedication to hedonism to completely set convenience aside for pleasure of driving yourself.

Further, better public transportation isn't a practical substitute for many scenarios in the United States where these driverless cars could make great inroads (no pun intended). It's not at all uncommon for someone to live 20+ miles away from the nearest public transportation in the United States without being "rural." If you're already far enough away from public transportation that it's unusable, you aren't likely to move without at least considering whether a driverless car will ease the pain of your commute.

Further, there are plenty of reasons that public transportation simply isn't a feasible option but a driverless car would rule. Two of my very good friends got married this summer, for instance, and their wives were from extremely rural areas such that plane tickets were ~$700 a piece. Instead of buying ~$3000 worth of plane tickets for my wife and I to attend each wedding, we drove 800 miles to each for far cheaper. If I could tell my car on Saturday night "Take me to Peioria, IL - and I need to be there by 3 PM on Sunday. Pace yourself accordingly." and know that I'd get there on time and safely, I'd absolutely do it.

I think that, within 20 years, driving manually will be seen in a similar vein to how driving a stickshift is seen now, and within 50 years, it'll be unusual for anyone to drive themselves anywhere.

>> Eating food from a tube could also give us a lot more free time but we don't do it because it is not an enjoyable experience. We not only enjoy the flavors and textures that regular food provides, we also enjoy the rituals that accompany eating. In this regard driving cars is somewhat similar.

Yeah, I felt this way for about two weeks after I learned how to drive. After that, it quickly got tedious and boring.

I've been driving for over 10 years and I still enjoy almost every journey - I think autonomous cars are a fantastic idea but I would miss driving if I couldn't ever do it.

I don't think the automation would ever be mandatory. It would be more like auto-pilot, where it would do most of the mundane driving, but let you take over when you want.

Do you think once the automated car gets to the point where it causes 1000x less fatalities, that people would still allow the freedom to drive a 3 ton missile around public? I personally think not.

The first software recall on the driving software might be the end of it. Its going to take a while, probably a generation or two of people.

It was never enjoyable to me, what with moving tons of steel 5x faster than primates evolved to move at. And all the other people on the road doing the same thing, often with texting and coffee and other distractions making it all the worse.

I can't wait for the self driving car, because I have much more faith in computer systems than your average person off the street.

> Eating food from a tube could also give us a lot more free time but we don't do it because it is not an enjoyable experience.

Even if you thoroughly enjoy driving, a driverless car would afford other emotional highs from:

- Allowing you to get an extra [insert unit of time] of sleep - A bit of hanky-panky (even the one-man party variety) - Eating a meal (preferably not from a tube) - Playing games (a driving simulator if so inclined) - Reading HN

I love driving as well but I can see the benefits that'll come from automated cars will easily outweigh the negatives. One of those positives might be a more energised, happier workforce.

"but we don't do it because it is not an enjoyable experience."

Eh, no we don't, we don't do it because it's not available on the market and because there is no way to make such food sit well with our digestive systems, otherwise many many people would 'eat' (or should I say 'drink' or just 'swallow') food like that. I'd pay good money for not having to think about what to cook, actually cooking it, doing dishes etc. 3x7 times a week (if I still had the option to eat recreationally - i.e. no solutions that replace my digestive system with something else allowed).

I'd eat food from a tube. I don't enjoy the flavors, textures, or rituals that accompany eating. (Sometimes, I even have a caffeine pill instead of a cup of coffee in the morning. Gasp!)

Some exceptions do exist, like driving in heavy traffic

Which is a large "exception", since by definition lots of driving occurs in heavy traffic.

> Autonomous cars are the next big tech that changes the world - I put them up there with computer, Internet, and smart phone.

I wouldn't be so sure. Yeah, it might prove to be a very good thing for the United States, but for the rest of world there's first Europe's public transport networks (I'm 31, don't have a driver's license, I live in Europe, I'm doing perfectly fine, thank you), and short of us reaching the Singularity I honestly don't see any of these cars driving in the crazy traffic of Cairo, Istanbul, or any other metropolis from the rest of Asia or Africa.

It's going to shift a lot of capital that's currently tied up in owning and maintaining cars into other uses. Right now, most cars are only in use for about 5% of their lifetimes. When everybody owns their own private car, those cars spend a lot of time sitting parked. If those cars could drive around and be used by other people who need them on demand, you'd need a lot fewer cars, as well as a lot less parking.

There are so many things I can think of. The taxi/rental/on demand model is far more likely to happen because who needs to own a car 100% of the time. Ride sharing is also easier - I suspect many people would be happy going a few minutes out of their way for halving the cost of their journey. (Computer systems would automatically manage supply and matching demand, as well as pricing.)

Not only do you need less parking for fewer cars, but you could have less parking for the same number of cars. For example there is no need for stores to have huge car parks - customer and employee cars can go and wait somewhere else and only return to the front door just as you come out with your goods. This can lead to far denser human centric areas.

Some roads can actually be decommissioned. Vehicles can load balance across existing alternatives and respond far quicker to developing bottlenecks. Occupied vehicles can be given priority over shorter routes than non-occupied ones.

Deliveries can be done without drivers. For example the delivery vehicle can drive to your home, work or anywhere else and you enter a code to get your items vending machine style.

It will give people greater freedom. For example if your vision isn't acceptable, or you have physical issues then it won't matter. In general the maladies that come with age won't affect mobility anywhere near as much as they do today.

Drives can be over longer distances as there is no driver fatigue. Perhaps there could be sleeper cars for when you want to go somewhere overnight.

Heck they can even do things like deal with you in the vehicle while you are driving to an airport or a border. You could show your passport, answer questions, check luggage etc. And then go straight through the border or walk straight to your gate at airport, or even be driven to a closer access point to the gate.

And I suspect we never get flying cars out of this. When the two dimensional earth surface is used so much more efficiently, why expand into 3 dimensions? And since we can increase the speed the more vehicles are automated, what advantage does flight have?

Remember that the "5%" of the time you use your car is probably the same "5%" of the time every one else wants to use their car, i.e. rush hours.

Yes, it seems unlikely that the same car could be re-used for multiple trips. However, you could move the parking lots far away from the places people want to go because the car could drop you off at the door and then valet itself to a remote lot. That would at least free some of the very valuable space currently used by downtown parking lots.

Nobody is talking about replacing a car part of let's say 100 with a car park of 5. If we could just replace it with a car park of 50, which with current travel patterns is perfectly reasonable, that would be a tremendous gain.

The "lot less parking" will do wonders for urban landscapes. When I travel through cities (including my own) and land that could be used for parks or habitation swallowed up by parking space, it seems like an incredible waste.

Though this presents a problem in that we're burning through a car 20 time faster, (100% vs 5%). Most cars people own don't last 20 years, so they won't even last a year in the future? There better be some distributed economic system for providing those cars, because that's a large expense.

A car right now needs regular maintenance as it is every few months, even if it's just changing the oil. That will go from months to days if they're being used 24/7.

Actually a car that is driven frequently holds up better than a car that is only driven every now and then, and is left outside for the rest of the time. Cars in almost-constant use will also be taken care of better, and automatically; for example, they will be driven through a car wash automatically regularly, and the automatic oil monitoring system will replace said oil at a machine during 15 minutes somewhere at a point where the demand is lowest. Maybe we'd be burning the car 2 times faster, I don't know exact numbers - but it surely won't be 20 times, and for any multiplier lower than those 20, we come out ahead.

Carsharing systems are already dealing with these problems successfully.

> I'm 31, don't have a driver's license, I live in Europe, I'm doing perfectly fine, thank you

I have a license but don't have a car (actually owning a car was too much of a hassle) and you're only telling half the story. Yes, day to day not having a car is fine, but every once in a while you want to buy something larger that is not delivered, or move some stuff, or go for a short vacation, or even just drive out of the city to pick some mushrooms and you're stuck. You either borrow/rent a car or skip on the activity.

A car does provide a kind of freedom that public transport simply can't. A car that drives itself, or, even better, a car that I can cheaply rent that will drive itself to my doorsteps, then drive me wherever I want to go, and then return itself would be absolutely perfect.

A car that drives itself, or, even better, a car that I can cheaply rent that will drive itself to my doorsteps, then drive me wherever I want to go, and then return itself would be absolutely perfect.

We have those, they're called cabs/taxis ;) Well, I suppose they're not exactly cheap, but personally I use them so rarely that it doesn't really matter, even though I'm on a tight budget.

Ubiquity and response time matter. I live in Los Feliz, a popular LA neighborhood full of lots of 20 and 30 somethings with disposable income. It's nigh impossible to hail a cab here. In fact, outside of NY and the downtown business districts of a few other major American cities at certain times of day, "cab on demand" basically doesn't exist. If I want to go out without driving on Friday night, I use an app like Taxi Magic to request a cab, then wait somewhere between 10 and 30 minutes for a cab driver who speaks relatively poor English, has never heard of the bar I'm going to, needs me to provide turn-by-turn directions the entire way, misses our side-street repeatedly, and then is annoyed that I'm paying with a credit card.

Imagine instead being able to use an app where, in any urban area, you can have a car sitting in front of your house in 3-5 minutes, with reasonable mileage rates, exact GPS coordinates preprogrammed, and no haggling over tip and payment method at the end of the ride.

Speaking as half a couple paying $1000 a month in car payments, car insurance, maintenance and gas, this is going to fundamentally change the calculus for private car ownership in America.

>Ubiquity and response time matter.

Yes. cabs are useless where I am, too. They say they will be there in 30 minutes, but that's best case.

But there are plenty of unemployed folks with cars. How much of the cab shortage is due to regulatory capture?

Imagine, for instance, an app where I could give my location, a timeframe and, say, an amount I was willing to pay. "Hey, $50 if you pick me up from the office and drive me home... if you get here within 10 minutes"

Such an app is technically easy, and I imagine there are enough people with cars either going where I'm going or simply idle to make it work. But there are... legal barriers in place.

Isn't Über in LA?

A (folding) bicycle can solve a lot of those problems.

Not having to own a car opens up more options

Want a big car for a family trip? Just order one and it turns up where you are, ready to take you where you want to go.

A sleeper car for two? A flash sports car for the afternoon? A van for the day?

Pay-as-you-go personal transport from anywhere to anywhere where you can still be productive during the journey... Sounds pretty revolutionary to me.

This idea already exists as car sharing schemes, it is incredibly handy. Self Driving cars will only improve car share schemes.

See the following sites for examples of car share schemes.



What if you join Zipcar to Ridejoy to Uber to autonomous cars? Half taxi, half bus, totally convenient.

"(I'm 31, don't have a driver's license, I live in Europe, I'm doing perfectly fine, thank you)"

Anecdote does not data make. I'm 32, I have a drivers license and two cars, I live in Europe, and my quality of life would be severely impacted if I didn't have a car. The vast majority of Europeans live in places where the population density is too low for it to be sensible to depend solely on public transport, not to mention that only a part of Europe has any 'reliable' (I use that word loosely, because name me one public transport system outside of the German one that works 'reliably') form of public transport at all.

It only takes one or two times a month that you need a car (for big groceries, visiting a far away friend or relative, or going somewhere where there is no public transport) to make it worth while to get a car, and once you have it, it's better to use it as much as possible to make it pay for itself. Public transport is not a viable long-term option to satisfy our transportation needs - it's too expensive, inconvenient and inefficient. Lean, pay-per-use and easily rentable autonomously driven cars are the way of the future, and the near future, at that.

Cars will become more appealing than public transportation over time. As the downsides of cars fall away, their packet switched functionality will ultimately win out, particularly for wealthier people.

If driverless cars eliminated the need for public transport (in the contemporary sense) you could open up rail corridors to create highways for cyclists, joggers, and walkers - it would open up alternatives to motorised travel.

Or you could turn the theory on its head - use the rail corridors as motorised transport arteries and open up the roads to the bipeds.

(From Sydney where the rail network is extensive - I'd imagine places with less ubiquitous rail wouldn't benefit as much)

I would think the reverse. Why own a car if you can get a taxi at your door in the size/shape/color of your choice whenever you need one without having to pay for the taxi driver?

A taxi is a car.

You are right, but I am, too :-). In my mind, a taxi also is public transportation. I do not think I am alone in that. Wikipedia disagrees with it in http://en.wikipedia.org/wiki/Public_transport ("as distinct from modes such as taxicab"), but http://en.wikipedia.org/wiki/Taxicab talks of "modes of public transport" for taxicabs. I guess the terminology will become more confused if/when driverless cars become commonplace.

> I honestly don't see any of these cars driving in the crazy traffic of Cairo, Istanbul, or any other metropolis from the rest of Asia or Africa.

You're wrong. Japan is also planning on introducing self-driving cars shortly.

I think the parent poster meant cities where the road rules amount to a functioning car horn. The roads of Japan are nothing like Cairo/Istanbul etc. There are a lot of cars - but they drive in an orderly, predictable manner.

On a related note: http://en.wikipedia.org/wiki/Yurikamome

In Tokyo they already have an automated transit system that is basically a truck on rails. See this image to get a better picture: http://en.wikipedia.org/wiki/File:Model_7000-Fourth_of_Yurik...

Yurikamome is just a driverless train.

It's not a train in the classical sense - it has rubber tires. More like a truck confined to a rail.

Either you have the perfect commute (rent must be expensive?) or you're just not paying attention to how much time you're using in travel. Public transport takes longer than driving in virtually every possible situation. The only equalizer is if your commute were near perfect, and short you might gain a slight advantage since the driver has to park.

But if we go to robot cars we no longer have to park.

Also, places like Cairo are bad because of the human drivers. If you only allow robot cars that all goes away. The system can know where everyone is and where they are going and adjust speeds to the point of having your car end up 20cm behind the next car instead of having to use the breaks if you went 1/kmph faster for the last 200km.

I'll tell google to stop work because it doesn't help with your daily commute :-)

So will susceptibility to car sickness become a federally-protected disability?

I wonder how much of car sickness is thanks to the driver. And if a good enough software could drive without causing car sickness?

Car sickness tends to happen when you're focusing on the interior of the car. That's what leads to the disconnect between your visual field and your mechanical sense of motion. I expect this is the main reason that drivers are rarely affected -- they're always looking out the window (more or less), so their visual sense of motion remains aligned with their other senses.

In my personal experience, car sickness is caused by high jerk (http://en.wikipedia.org/wiki/Jerk_%28physics%29), primarily during deceleration. (This is also commonly known as "bad driving".) I imagine this is something that self-driving cars won't have a problem with.

Seconded. I can take road trips like crazy, but hop in the Mountain View Caltrain shuttles with their crap drivers on a warm day, and I seriously feel like I could lose my lunch.

That doesn't necessarily account for people getting car sick on long, smooth, winding canyon roads driven at constant speed. The jerk induced by following the curves in the road should be lower than that caused by braking.

Honestly I think it is a combination of where your awareness is focused, and jerk. Long windy roads put varying lateral forces on passengers which probably become an issue if their focus is in the car.

Sometimes I get "carsick" while working with very small components (such as eyeglass-sized screws) that aren't behaving as they should, so I would have to attribute greater weight to the idea that carsickness is caused by a disagreement between the senses.

Well, at least in my case, it only happens as a passenger. I can sit there, and I'm fine, but the instant I try to read something, watch a video, or something like that, I start to feel bad.

Actually, even a glance at my GPS while driving wil occasionally cause a twinge.

I don't necessarily think that will be the case. As far as I knew (and from re-skimming the wiki article), car sickness (and any kind of motion sickness) is caused by the brain seeing different things than what it's processing from the inner ear. http://en.wikipedia.org/wiki/Motion_sickness

So it's more up to the individual in question, then the driver (considering a normal, average car ride, not anything in the extremes).

Maybe if you had some sort of AR setup that projected what you were interested in onto the environment. E.g., passengers would be able to read a book as if it was on a billboard floating on the horizon. Would that be enough to combat the effect?

In theory, yes. Motion sickness depends on the type of motion. A very bouncy or aggressive ride will result in more sickness, whereas a very smooth ride (such as a train) results in less. It's all about how much confusion there is between what you're seeing and your sense of balance.

As it happens, the same types of driving that are more fuel efficient cause less motion sickness. Still, driving anywhere with stop lights and other traffic probably necessitates enough stopping and starting that even maximally efficient driving software would cause motion sickness in many people.

For me, it is a side effect of some problems with my vision. I hear it is curable with vison therapy, but that is quite expensive.

That's actually really interesting.

I've always worn glasses but my vision has gotten progressively worse as I've aged - it's not bad, even now, I can legally drive without glasses - but it's enough to be noticeable.

I also was much less prone to motion sickness when I was younger - I would happily read full novels while riding along as a grade schooler.

My eyes have been awful since I was a kid, but just near-sighted. My daughter needed vision therapy as a young child, and when I went into the therapist's office, I got an exam for my glasses prescription.

Turned out that I had similar issues, just not noticeable. It had to do with my eyes not being able to properly maintain a focus at a specific distance over time, and their ability to track properly. I've always been a slow reader - this is at least partly why. I was told that it's also responsible for the fact that I get car sick within seconds of reading in the car, and that it goes away with therapy. Unfortunately, I never did go through with the therapy.

While this is very encouraging, the TC article is also very misleading. For example: "There have, of course been some accidents that involved Google’s self-driving cars in the past. All of these, however, happened while humans were in control of the cars." - isn't this far more likely to be the case, since the human only takes control when the car can't figure it out?

Also, who's commute doesn't involve temporary construction signs?

Good points. Without any statistics about the frequency and nature of situations where the human driver still has to intervene, numbers like '300K miles without an accident' are meaningless. For all I know the Google cars might have had a near-accident every 10 miles if it weren't for the human driver that's always there to save the AI when it fails. The fact that Google never discloses anything besides mile counts, and only lets outsiders take a ride with the Google cars under extremely controlled situations, is quite telling. Statements about Google car achievements like this are not much more than propaganda to me, until Google starts to share actual in-depth details about the performance of the cars. If it really works as well as they claim it does, what can they lose by disclosing everything?

Myself, I'm extremely skeptical about self-driving cars, and I've never understood why I seem to be massively outnumbered by people who are utterly convinced it is possible to create AI that can handle the billions of different and sometimes completely arbitrary traffic situations you'll encounter while driving on current road infrastructure. Sure enough it will work great 90% of the time in some places, but that's not enough. It needs to work 99% of the time everywhere. Simply looking at the achievements made in AI in the last 50 years or something doesn't really warrant a whole lot of optimism we'll get there anytime soon.

Would you be ok with a computer that augments your own driving (similar to those cars that automatically brake when they sense an accident)? If so, then it's just a matter of continuously improving on this augmentation.

Even if it never reaches the level where you can truly step away from the wheel while the computer drives 100 miles in rush hour traffic, attaching an AI to help avoid accidents in many common situations still sounds revolutionary. I forsee something similar to a plane's autopilot -- where it greatly reduces user error yet sometimes still needs a human to take over when it doesn't know what to do.

That's exactly what I think will be the future: AI-assisted cars, not AI-driving cars. But I don't think it will progress beyond that, for the simple fact that I think it's either all or nothing. You cannot build a car that drives itself sometimes, but still needs someone to take the wheel every now and then. People will stop paying attention to driving which will inevitably lead to accidents when they have to take over from the AI. It more or less defeats the purpose of a 'car that drives itself' anyway.

Don't get me wrong by the way, I'm not 'against self-driving cars' or anything, I simply think it can't be done. Not unless we replace all road infrastructure with something that supports autonomous vehicles, and introduce some form of centralized traffic control, ie: more like air traffic.

We are being promised a self-driving car. Either it drives itself, or it doesn't. If it doesn't, it's not what it claims. And if it does, what are the driving aids for?! That's like wearing a hat indoors.

Who promised you that? To me it's not a binary yes/no on whether this project is successful. Even a marginal reduction in the number of accidents will save lots of money and lives -- even if you still end up behind the wheel helping it along.

> All of these, however, happened while humans were in control of the cars." - isn't this far more likely to be the case, since the human only takes control when the car can't figure it out?

That's a really good point, all though I'd wager a guess that most catastrophic accidents (e.g. crashes at highway speeds and t-bones at left-hand turns) happen when the driver is lulled into complacency by normal conditions.

It's probably a bit misleading.

The car is pretty eye catching: http://www.vibe.com/sites/vibe.com/files/styles/main_image/p...

So it gets attention...and most accidents happen when people don't pay attention.

Once these become more common, it won't be a huge deal, so people will stop noticing them and accident rates will go up. Probably not by much, since you'd eliminate most of the driver caused accidents

You are suggesting human's will run into it more often right? I sincerely doubt we will see too many autonomous-car-at-fault crashes. The blacked out Lexus SUV's I saw at a Google event a few weeks ago were very discrete minus the velodyne sensor. If it adds an credibility I was on a team at the Victorville final in 2007. :)

As an aside, I've been waiting to buy my Google car for 5 long years. During this time I've become so disenfranchised with day to day driving that I avoid it at all costs. Commuting via roads in a car is such a mundane task that is, in my opinion, a solved problem. Bring on the autonomous cars!

Commuting does kind of suck if you're trying to roll in during rush hour. But without super-heavy traffic I find driving to be rather enjoyable in itself. Even more so if you're riding a motorcycle :) Perhaps the problem is not commuting, but the fact that many companies want their employees at work at 9:00 and out of work at 5:00?

The problem for me isn't commuting, it's having to invest so much money into a huge pile of metal and plastic that someone distracted by someone or something in their own pile of metal and plastic can destroy in an instance, possibly killing me and anyone with me. Or maybe I get distracted and the result is similar.

I just don't see the few cases where it's enjoyable to be worth the other costs. I'd consider it similar to smoking in that regard.

Feel free to refrain from driving, in that case; just don't get in the way (in a legal sense) of those who still want to.

Smoking is a much greater nuisance than driving, because a nearby smoker will foul the air for a fairly large radius through no choice of the people around them. Driving only affects other people who choose to be on the road.

Comparing to smoking seems to be a huge stretch to me. One thing virtually guarantees cancer and an eventual unpleasant death while providing no measurable benefits at all, and the other allows you to easily travel all over the place cheaply with a relatively small chance of a major accident.

virtually guarantees cancer

That's not true at all, the incidence rate of lung cancer in smokers is less than 20%. That's still much higher than for non-smokers (0.4% - 2%), but hardly a "virtual guarantee".

Will you be OK with google tracking everywhere you go and serving you ads for every shop you pass?

My wife sees a neurosurgeon who's an eight hour round trip away. I've driven her there and back at least six times this year.

With a Google car, I'd a) be able to send her by herself and b) even if I had to come, I could work on the way, meaning I'd have more than a week's worth of vacation/sick time restored.

Yes, I'd happily accept ads.

You could also hire a limo. I suppose in the future you'll be able to rent a google car for a day, but you don't have to wait for self-driving cars. They're here today, the self just happens to be a human.

That's not really reasonably cost effective for most people.

Having a full-time driver on retainer 24x7 is not cost effective, but using one a few times a year might well be.

For an 8 hour drive? Still not true, especially with a limo.

Neither is a Google car, though. Adding self-driving requires five figures, probably even in production.

I think $10,000 is a big over-estimate, but even that would be spread out over the life of the car. I'd happily pay $1000 a year to not have to worry about driving.

I'd happily pay $1000 a year for a car that doesn't drive itself.

A limo with a driver is always going to be fairly expensive. A self-driving car should come down in price a lot faster.

Hopefully I'll be wearing my google glass too! So, I guess so.

True privacy is very hard. You could set up a camera to read plates at some choice points and know the whereabouts of most of the people in your city. If you want to do something privately you have to act deliberately in this day and age.

Setting up those cameras would be illegal where I live.

Whereabouts is that? I hear many people make those claims, but almost as many turn out to be wrong when investigating more closely.


To set up such cameras, you'd need permission from our National Data Protection Commission and if you were to store footage with car plates you'd fall under the Personal Data Protection Law which requires you to either get explicit consent or fall under one of the few exceptions in the law (and no, tracking random people isn't one of them).

And even if you qualify, you can only use that data for the specific purpose you said you needed it for and it needs to be well protected (for example, a governmental CCTV installation was suspended because the recording weren't being well encrypted).

If you have an android phone, Google already has the capability to do this.

I believe they do, as well. IIRC that's how the realtime traffic data comes in.

I'll take that over feeling like I'm about to die while a careless driver is at the wheel any day.

Yes. I'll be asleep.

You don't really need ads for the shops you're already passing.

Good evening, John Anderton. Routing you to the pre-crime facility for detainment.

I see you're on your way to the drug deal you briefly discussed over gmail, i've taken the liberty to ensure the local human authorities are aware.

Better: robot authorities. Who can be bought off with a few bitcoins :)

On the other hand, staring at something is a great way to hit it. It's one of the first things they teach you in a motorcycle safety course.

The cars still need to learn how to handle snow-covered roads

That is going to be hard to pull off. For places that get a lot of snow, often you can't even be sure if you're on the road or not.

In theory autonomous care would be able to include a number of other factor to decide if they are on the road or not. Including, GPS, wifi signals, the grade of the road, the texture of the road, etc.

What about situations where lanes have been plowed off-center? Even the most accurate GPS and accelerometers aren't going to help the vehicle understand that the appropriate path has been shifted.

The car would still factor in the 3d analysis of nearby objects so assuming the use technology that works in heavy snow/rain fall they wouldn't run into anything.

A snow-covered road is a situation where:

  1. Road indicators are difficult to identify
  2. Roadway has less traction than usual
  3. Humans are having emotional reactions (joy, terror) which impacts their processing abilities.
3 is mitigated by design. 1 is a sensor issue, it's also a problem that very accurate GPS can solve.

There are many causes of slick road surfaces, from snow and ice to excessive water, spilled materials, collision debris, and low-quality roadways (dirt and gravel roads). Driving in reduced-traction situations shouldn't be an edge-case.

Actually the problem with snow is specific to Google's approach and is not limited to snow, but to all changes of the environment. They use the 360 degree LIDAR (Velodyne) and drive multiple times over each street in order to create a 3D map which is used for localization after that. Therefore the snow changes the layout of the environment significantly and localization stops working, which in turn makes functions like lane keeping, car detection, traffic light detection etc impossible. This will also happen if the environment changes a lot in another way.

This brings up another question: How will these cars react when a street is blocked and unpassable, due to heavy snowfall (unplowed), heavy rainfall over the street, mudslides & rock slides, etc?

What about where a street has parked cars on both sides and just enough room for one car to go through the middle of the street - one of the narrow city streets where drivers have to wait in the intersection for oncoming cars - will these autonomous cars wait and let drivers through, or will they get stuck?

Personally, I can think of lots of edge cases that will be interesting to see answers to. For instance, a 4-way stop - we know how to handle that easily, and can overcome someone jumping out of turn, again, quite easily.

More importantly - who will be maintaining all the traffic changes? An example - a section of Interstate or road is closed for months, whether by damage or construction. And then, when it's clear again?

I'm interested in seeing how these things play out.

The 4-way stop was something they had to change from the ideal setup. The google car was obeying the law and waiting its turn and never getting it. They made it get more aggressive the longer it waits and now it gets through them just fine.

Long-haul tractor trailers. We need long stretches of highway where autonomous tractor trailers can carry goods. It'll provide a good business case and limit liability.

We have (one of the) largest freight-train networks in the world. It's great for things like that.

Trains are basically semi-autonomous, and out of the way of cars anyway.

Our freight network is great, but trucks are now a vital partner to trains and go more places. If all the trucks in the US stopped and only trains ran, we'd be in a serious world of hurt.

It would be nice, but I can see the reaction from truckers who become out of work already...

Aren't autonomous cars required to have a human "driver" while on the road? I can definitely see truckers lobbying to have some sort of law of that nature - they sit behind the wheel "in case something goes wrong", but would be allowed to drive longer hours with fewer mandated breaks. Sounds like a win for the profession in general... setting aside that longer drive times mean that at least some will be out of work.

Even if the truck's "supervisor" is paid as much as drivers are today, trucking companies are very hungry to get more miles driven in fewer days. These will let trucks drive for more than 11 hours, thus I believe we'll see these as soon as they're approved.

They do now. Doesn't necessarily mean that in 5 years they still will.

Do unions still hold significant political sway these days? I can definitely see truckers pulling their weight to keep those laws on the books for commercial freight at least long enough to stay employed 'til retirement.

Assuming the laws are removed at some point, there would still be a huge transition between companies needing truck drivers and not needing truck drivers.

They could just stop hiring new truckers (forgot what it's called when you just let the current staff dwindle, instead of actively firing them) and let the current ones gradually retire. Either way, this is a problem that has been faced before countless times as new technology has appeared.

This was the easy part.

The hard part will be fighting the technophobes when somebody drives into the side of one of these and kills themselves. I don't envy being in that battle.

I would bet they can reconstruct the entire environment through their LIDAR records, play by play, and show the events leading up to the collision.

I hadn't even thought of that. I wonder how this will effect the car insurance industry.

I don't think this will be a problem because we're likely to see a gradual move towards fully self-driven cars where humans still drive but the computer takes over only when needed to avoid accidents. Volvo already has a system that stops you rear-ending other vehicles and BMW has a system that safely pulls the car over if the driver suddenly gets incapacitated. Eventually the most advanced cars will never cause an accident while still being human-driven. From there the leap to let go of the wheel is not that big and I think it will be reasonably easy.

I'd love to live in a world where roads were full of auto-driving electric cars that automatically went to recharge and were replaced by fleet cars, and stopped at various points for people to load up. Like personal, on demand busses.

> stopped at various points for people to load up.

By that point, we will all have integrated computers in our body, probably with some kind of brain-machine interface so we don't need to interact with them. We can just telepathically request a vehicle to take us somewhere.

Screw that, I want Star Trek style beaming up by the Scottie5000 bot in space.

> Like personal, on demand buses

That use a heck of a lot more gas ;)

No reason they couldn't be electric. One of the big barriers for electric cars is a short running length and long recharge cycle. Neither is an issue if your 'car' is just an (autonomous) taxi service that runs off to recharge itself when you're done.

Obviously that doesn't mean it isn't using a fossil fuel somewhere along the power chain, but I imagine the efficiency isn't much worse than bus routes - and it's not like they need to target the exact same use case, either.

Hence my use of the word 'electric' :]

It won't be consumers who drive the adoption of robot cars, it will be the logistics companies. This will all start with factory-to-doorstep supply chain integration.

First, Amazon and Wal-Mart and Fed-Ex will have giant fleets - then consumers and mid-sized businesses will follow. Small local governments and developed Asian nations like South Korea and Japan will also drive adoption.

Having a fully-automated supply chain will break the competition who doesn't have the capital to get there fast enough. The big companies already automate the factories, and once delivery is automated, the circle is complete.

>Google warns that “there’s still a long road ahead.” The cars >still need to learn how to handle snow-covered roads, ...

What if the car was smart enough to identify situations where it was unable to auto-navigate safely? It could pull-over and ask the human to take-over for awhile. I would rather see a mostly capable self-driving car come to market sooner than wait around for near perfection.

On this note, it seems that the trucking industry could adopt a self-driving truck even if it was only good enough to travel along long stretches of highway. The first/last miles would be manually driven just like today. Could the trucking companies use remote navigation (much like military drones) to handle first/last mile and perhaps problem spots along the way?

What's the minimum viable product?

> What if the car was smart enough to identify situations where it was unable to auto-navigate safely? It could pull-over and ask the human to take-over for awhile.

This is precisely what the Google car does, in fact.

Does anybody have a citation for accidents per mile for a human driver? I've seen, uncited, 1/200,000 but when I tried to verify this I came up with 2,967 billion miles traveled in 2010 [0] and 5,419,000 crashes total (based on police reports for: property damage, injury, or fatality) [1]. Which, unless there is an error in my math, is roughly 1 crash every 550,000 miles.

0. http://www-fars.nhtsa.dot.gov/Trends/TrendsGeneral.aspx

1. http://www-nrd.nhtsa.dot.gov/Pubs/811552.pdf table 4

I'm of two minds on these.

On the one hand, these will help close the "fun gap" between the city and the suburbs. One big problem with living in the suburbs is that you can't really drink a lot wherever you are, because you have to drive home afterwards. If your car drove itself that restriction would be removed, and over time the suburbs could start to compete with cities for nightlife despite the large delta in density. Places might even start staying open later because they could make more money selling alcohol.

On the other hand, I have a power-nothing manual-everything sports car that I love to drive. Already most people I know drive cars with automatic transmissions, and now some manufacturers don't even ofter manuals anymore, even on fairly powerful high-end cars. I can imagine many of the same people who choose an automatic transmission today will choose a fully automatic self-driving car tomorrow because to them it's a utility and not a form of entertainment. Eventually, the market for non-self-driving cars will shrink in the same way the market for cars with manual transmissions has.

The problem is, if it gets to production, the first real accident will be considered totally inacceptable.

It doesn't matter if you reduce by a 1000 the risk of accident. When an accident is caused by a human error you have someone to blame, you can think "I would have done better". But when the accident is cause by a machine, then people will stop trusting the car and be done with it.

I've seen this argument made in many forms. The short response is: "Google already thought of that."

People will be upset, even outraged, by the first few autonomous car accidents. But as long as they're still legal, they'll be building up a safety record that will quickly surpass what humans can achieve. Google has already invested enough money in the legislative system to prevent laws from being yanked off the books after one accident.

The general form of "People won't like it when machines do X" doesn't have a lot of merit. There are always problems and people never like it at first. Then they get used to it and quality of life improves over time. I'm sure similar arguments were made about the automobile itself.

Paraphrasing Peter Norvig: ".. if you do experiments and you're not failing half of the time, then you aren't getting enough information out of those experiments."

(from http://www.slate.com/blogs/thewrongstuff/2010/08/03/error_me... )


I wonder what kind of denial of service attacks we'll see on autonomous cars once they're popular.

Trivial example, could i choose to always pull out in front of the robot even if it's not my right of way? (presumably safe in the knowledge the robot will always give way to avoid a collision)

Could that escalate to the robot being treated as a mobile speed trap / felony snitch?

Interesting times.

I kind of wonder how Google's driverless car handles the situation in extremely populous cities such as Shanghai or HongKong.

That's part of what I'd like to see more.

  - scalability cases : many many moving obstacles.

  - reversal cases    : gCar behavior if a vehicle is misbehaving. 
                        What happens ? can they compute a safe avoidance path ? 
                        if not but with impact still being detected in advance 
                        by the system, can there be a pre-protection system ? 

  - systemic cases    : how 2 gCars react to each others, any oscillatory patterns ? 
                        computing a path using the same heuristic leading to 
                        avoidable collision. With more than 2 ?

Forgive me, but could you reformat that? In Firefox, the code-ified view of your text is taxing to read. It has an unpleasant scroll bar and I forget what I've read by the time I get halfway through it, since I can't see it.

Right, made it a fair bit more readable now. I'm difficult to read anyway.

Much better, thanks!

In the city, when there're many pedestrians on the roads, I would expect from autonomous car to drive slower and stop in case of pedestrian spotted (human drivers are smarter, but they usually violate speed limits and have worse reaction time)

I'm afraid that in New York, this would just lead to total gridlock as a few people would stop traffic by jaywalking, leading to a bigger flow of people following them.

There's a classic science fiction story that I cannot remember now where people do exactly that.

If it becomes a problem, the car can transmit images of the jaywalkers to the police.

I'd love to see it handle the 17th century layout of central london as well.


I see them in heavy traffic on highway 101 near Mountain View most days of the week.

Important detail: "It’s not clear how many of these 300,000 miles were driven on Google’s secret racecourse, by the way."

I would love to see google incorporate the efforts of SARTRE, a system of caravaning cars on highways for reduced drag: http://www.sartre-project.eu/en/Sidor/default.aspx

I wonder how well these cars handle heavy snowfall or ice on the roads. I know firsthand how driving becomes infinitely more complicated under these conditions and it would be major selling point for these cars if they could handle it smoothly.

They don't handle snow at all yet. There is no question that will be quite a hurdle to cross, but I doubt it will be for the reasons most think. My guess is figuring out where the road is at is going to be the hardest part; where I bet most people think it is "how do I drive on ice". Without impatient, jacked-up-4x4 drivers driving a twice what conditions warrant, it is just a slightly different calculation of static and dynamic frictions.

Put another way, Google's Autonomous cars complete less than half of the total daily commute of a small metropolitan workforce.

That's not a lot of miles, but the fact that they've done it completely without accident is relatively impressive.

You will not know how safe they are until accidents happen. I would love to see someone with money run an experiment with real drivers vs google cars. Create a worst case scenario course with simulated drunk drivers, blown tires, missing stop signs, etc.. and run this with real and fake drivers. I suspect the average driver would crash 10% of the time, autonomous car would crash 1% of the time, and a professional driver would crash never.

The real safety improvements will come when all cars are autonomous and they talk to each other.

"Missing stop signs" is one place where a google car would be great, since they pre-map the area and would know a stop sign has disappeared. Knowing that would cause them to slow down and be even more careful.

Autonomous cars never get distracted, have 360 degree vision, look at all things all the time, and can react in tens of milliseconds versus hundreds of milliseconds. I'd trust that over a "professional driver."

I doubt that there were really no accidents at all or close calls, but even if that is the case, you have to expect some accidents as more and more of these vehicles get out there. But even if there are 1000 accidents in the next 300000 miles or whatever, I hope people are realistic and give the technology time rather than writing it off as soon as a very well-publicized collisions occur.

In the future, there must be some communication between vehicles how to minimize the effect of the crash - e.g. if in one vehicles there is no-one, and no hazardous material - then this car somehow should be able to take more of the damage.

Oh, and I would very much like if every vehicle put on the back and front of the car - it's current speed measured by the car.


Applications are open for YC Summer 2019

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact