Volvo Testing Autonomous Cars On Public Roads

cartechboy writes: "Multiple automakers have already committed to selling autonomous vehicles by 2020, but only a handful of them have actually started testing and developing them. Now Volvo is putting self-driving test cars on real public roads in Sweden among other, non-autonomous traffic. 'The test cars are now able to handle lane following, speed adaption and merging traffic all by themselves, Volvo engineer Erik Coelingh said in a statement. 'This is an important step towards our aim that the final Drive Me cars will be able to drive the whole test route in highly autonomous mode.' The goal for the Drive Me project is to deliver 100 autonomous cars to customers in Gothenburg by 2017."

Comments

Google getting all the glory?

[mwvdlee]

How does this Volvo project compare to the famous Google cars?

Re:Google getting all the glory?

[Anonymous Coward]

How does this Volvo project compare to the famous Google cars?

Dunno but the kool-aid factor is definitely way lower for a Volvo.

Re:Google getting all the glory?

[timeOday]

Volvo is just now putting autonomous cars on public streets, so google is getting all the glory for a good reason - they are years ahead. Google had logged 300,000 miles by 2012, and now have reached 700,000 miles. And in contrast to highway driving like in this blurb (lane following, merging), google is way beyond that; highway driving is almost a given and google has moved on to construction sites, pedestrians, cyclists...

Re:Google getting all the glory?

[InsultsByThePound]

Soon google will give you the car for free, downside is the outside is covered with e-ink, and the future will look like one huge f1 race, sponsor wise.

Re:Google getting all the glory?

[ebno-10db]

And how many cars has Google ever sold? With $150k of extra equipment bolted on, they neither know nor care about the realities of car production. Volvo is intent on actually putting these things in the hands of real paying customers. To Google it's just another cool toy to play with. Hire some good AI people and have fun - who cares if it has nothing to do with their business.

More importantly, there are too many unknowns about the Google cars (despite them having been around for years), and about the Volvo cars. How do the cars perform in the rain? Snow (I'm betting Volvo thinks about that more)? How often does the autonomous feature kick out and require the human driver to take over? How much time does the human have to respond? In testing, how often do the human drivers take over control because they anticipate that the car is entering a situation where the autonomous control will have problems? Etc., etc., etc. Without that information we have no way of knowing if it will soon be a viable technology, or it's just a cool demo. However, Volvo is planning to sell these things in 2017, so they're probably working hard on a realistic approach. What year has Google claimed they'll ship product? Sergey's "5 years"? Sure, and I predict we'll have fusion power in 10. Honest. I hired some cool physicists who say they'll have it done by then.

Re:Google getting all the glory?

[PRMan]

Volvo just caught up to what Mercedes has been doing for 5-7 years already.

Re:Google getting all the glory?

[timeOday]

The reason google is getting all the glory is because they are going big and aiming for the moon - automated driving. The reason Volvo will beat them to market and turn a buck before google is because Volvo, like the other automakers who are working on it (Daimler, Toyota) are taking an incremental approach, and adding automation a little at a time. They want to make a more moderate investment and make money on it sooner, and who would blame them for it?

But google's long-shot may well pay off, too. If google can get just a few thousand highly-instrumented, networked cars driving around, they will have the world's most detailed and up-to-date multi-sensor maps of the roadways by a HUGE margin. This may enable cheaper cars with fewer sensors to be viable, because all they have to do is a sanity-check on the high resolution map and monitor obstacles such as cars and pedestrians. But they will download (on the fly) the location of every stop light, detour, pothole, school zone, you name it. By linking to google's data center, each car will be as familiar with its surroundings as you would be if you drove up and down the same street all day. Thus it is possible all car makers will end up paying google handsomely if they want a competitive automated car.

They will still be expensive at first, but getting critical mass is the hard part. After that, roads will be instrumented and detailed real-time maps will be maintained, so the cars won't need such high-end sensing capabilities in the long run.

Re:Google getting all the glory?

[newcastlejon]

So long as the inside isn't plastered with ads I really couldn't give a toss, but I think we both know that isn't going to be the case.

Re:Google getting all the glory?

[WindBourne]

Google has 100's of cars on the road for the last 5 years.
Volvo and others are JUST GETTING STARTED.
So, when Tesla gets Google's tech with minor changes to their cars, who do you think will be ahead?

Re:Google getting all the glory?

[WindBourne]

Google's algorithms will be added to Tesla, which is already considered far safer than anything that Volvo has/had.
And Volvo is not planning on selling these in 2017. They are planning on TRIALING these in 2017. There is a HUGE difference. Google will have theirs on Tesla, by the time that Volvo is just getting their first test vehicle out there.

Re:Google getting all the glory?

[Anonymous Coward]

So, when Tesla gets Google's tech with minor changes to their cars, who do you think will be ahead?

Uhm.. this is Google's tech

Volvo Car Group and Google Inc. said they have begun testing self-driving cars on city streets, a crucial new phase in the quest to make the technology a standard feature in automobiles.

Tesla might fit Google's tech to their cars later but Volvo has already done it.
I am pretty sure Volvo will still be ahead when Tesla gets around to it.

And regarding

Google has 100's of cars on the road for the last 5 years.

there is a pretty big difference.

"A mile of city driving is much more complex than a mile of freeway driving, with hundreds of different objects moving according to different rules of the road in a small area," wrote Chris Urmson, the director of Google's self-driving car project in the blog post on Monday.

Re:Google getting all the glory?

[dinfinity]

I'm betting that consumer acceptance will also benefit greatly from the incremental approach: get used to bits of driving getting taken over bit by bit is a lot less daunting than getting into a fully automated car and just pressing go.

In that sense, the automakers are doing the world of automated driving a bigger service than Google.

Re: Google getting all the glory?

[Anonymous Coward]

I live in gothenburg and i am pretty shure i read in the local newspaper that demo cars would enter the streets in 2017, not that they would start selling them.

Re:Google getting all the glory?

[Reziac]

And I think you just nailed why Google gives a shit: because a self-driving car that never takes a day off and doesn't demand benefits is WAY cheaper than a gaggle of employees driving those same miles. Since the object is to accumulate data to sell (to other SDC manufacturers in due course), why spend all the profits on employees if a robot can do the job cheaper in the long run?

No way I could trust a self-driving car

[MouseR]

when GPS still send us to wrong destinations.

Re:No way I could trust a self-driving car

[Anonymous Coward]

I've never had a GPS send me to somewhere that I didn't ask it to send me. The error people commonly report is in fact human error in entering the destination, not machine error in taking you there.

Re:No way I could trust a self-driving car

[Karl Cocknozzle]

I've never had a GPS send me to somewhere that I didn't ask it to send me. The error people commonly report is in fact human error in entering the destination, not machine error in taking you there.

Congratulations on only entering destinations that have existed for at least a year! Those of us living in "newer" portions of cities enjoy the constant joy of having our street "not exist" on Google Maps/Mapquest/wherever people are going now besides those.

Re:No way I could trust a self-driving car

[Todd Palin]

The GPS issue doesn't bother me, its the bad drivers. Unless ALL CARS are self driving the problem is that there are going to be lots of bad drivers doing stupid things for stupid reasons. A self-driving car probably won't be as good at anticipating stupid drivers as a really good human driver. There are stupid pedestrians, and stupid bikers as well. If there were no stupid, people self-driving cars might work. Have you ever seen a stupid driver try to merge onto a freeway with their turn signal on and their foot on the brake? Or see three cars bumper-to-bumper trying to merge onto a busy freeway as a pack? Can self-driving cars see this about to create chaos for the other cars?

Have you ever seen a biker ignore a stoplight because they think the laws don't apply to them? Will the smart car be watching the biker and anticipating stupid behavior?

What choices will a smart car make when an accident is inevitable, lets say it has a choice between a head-on and hitting the ditch at 65?

Re:No way I could trust a self-driving car

[Anonymous Coward]

I'm looking forward to the day when I can run people over and blame the car. It will be great fun.

Re:No way I could trust a self-driving car

[Anonymous Coward]

Self driving cars always pay attention, have a faster reaction time, and have fewer blind spots.

I have a perfectly clean record, and consider myself a good, attentive driver, but I'd trust a self-driving car any day.

Re:No way I could trust a self-driving car

[ShanghaiBill]

A self-driving car probably won't be as good at anticipating stupid drivers as a really good human driver.

Why should SDCs have to be better than a "really good human driver" rather than an "average human driver"? As long as SDCs are better than an average human, they are a net win. Also, bad drivers will likely be the first adopters, both because many bad drivers don't enjoy driving and because they pay higher insurance premiums.

Re:No way I could trust a self-driving car

[Anonymous Coward]

I think you are conflating a machine's decisiveness and predictability with intelligence.

If we could build machines which were smarter than us, there would be no us.

Seems cheaper and 'smarter' to build a better driver, as Germany and others do.

When people are bad at something, in this case driving, it is because they don't want[have] to be good at it or they are incapable[they should not have a license at all].

Re:No way I could trust a self-driving car

[pr0fessor]

You forgot the restaurant that's now something else.

Re:No way I could trust a self-driving car

[Anonymous Coward]

Have you ever seen a biker ignore a stoplight because they think the laws don't apply to them? Will the smart car be watching the biker and anticipating stupid behavior?

Sigh... you're missing the obvious. Here, let me spell it out:
Next thing you know, Bing will be making announcements for self-driving bicycles.

Steve Ballmar's successor will partner will a coalition of other automobile manufacturers who share statistics about how Microsoft self-driving bicycles typically result in less expensive damage than the average self-driving car accident. Where do you want to go today?
Could stupid behavior other than a biker's be successfully predicted? Sure, but then there's another potential invention just waiting to be made. The underlying rule remains: As long as stupid behavior exists, business opportunities will abound.

Regarding your last question: a very similar question was already asked on Slashdot. (Thread 1, Thread 2) You make it sound like a car needs a way to evaluate a head-on collision with hitting a ditch at 65. Which is more preferable? Spreading the pain so that multiple people get hurt, or making one person get hurt even more? What about the lawsuit when a car drives off a cliff because the car's automated driving thought that would be smarter, but really the car just caused an unnecessary accident because of a problem with the AI (or some other problem with the car, like a temporarily malfunctioning sensor)? The answer I came up with is this: this topic is way too undeveloped to have solid answers for all of these types of sticky questions yet. Fun times we live in...

Re:No way I could trust a self-driving car

[Carnildo]

I don't know about GPS units, but I've had Google Maps send me on a complicated route through the alleys of a small town because someone forgot to enter a permitted turn at the intersection of two major highways. I've had it tell me to drive through a concrete barrier because someone recorded the intersection as a cross intersection rather than back-to-back "T" intersections. I've had it give me a route four hours longer than necessary, because it thought part of the short route was still closed for the winter. And most recently, it give me a route that ended twenty miles short of my destination because it picked the park administrative headquarters in a nearby city as the location of the park, rather than somewhere actually, you know, *in* the park.

Re:No way I could trust a self-driving car

[geekmux]

The GPS issue doesn't bother me, its the bad drivers. Unless ALL CARS are self driving the problem is that there are going to be lots of bad drivers doing stupid things for stupid reasons. A self-driving car probably won't be as good at anticipating stupid drivers as a really good human driver. There are stupid pedestrians, and stupid bikers as well. If there were no stupid, people self-driving cars might work. Have you ever seen a stupid driver try to merge onto a freeway with their turn signal on and their foot on the brake? Or see three cars bumper-to-bumper trying to merge onto a busy freeway as a pack? Can self-driving cars see this about to create chaos for the other cars?

Have you ever seen a biker ignore a stoplight because they think the laws don't apply to them? Will the smart car be watching the biker and anticipating stupid behavior?

What choices will a smart car make when an accident is inevitable, lets say it has a choice between a head-on and hitting the ditch at 65?

With the number of stupid people driving stupid things that you've identified here that affect our ability to achieve autonomous eutopia, one can only hope that Darwin receives many more award winners to ensure our overall safety. Like any other time behind the wheel, either exercise common sense, or become a statistic. Your choice. Chances are they'll be programmed to handle some level of human error...they just won't be cranked up to handle Vehicular JackAss-assins on the road until at least a few generations in.

How many times did the stupid driver pull their stupid move and roll through their favorite light with the brand-new red-light camera installed? I'm willing to bet even the densest moron stopped after half a dozen tickets, so behavior can be modified. You just may not get a second chance to learn against a ton of steel.

Yup, downright cruel I know. Truth hurts.

Re:No way I could trust a self-driving car

[timeOday]

Google maps is far more accurate and up-to-date than Garmin and TomTom maps (the two that I've used), probably since more people use google and because google is more responsive in updating it. Yet clearly it's imperfect.

That said, staying on the road vs. navigation are two different things for the self-driving car, just as they are for you. The fact that a GPS thought you could take a route that is actually blocked by a concrete barrier, does NOT mean a self-driving car would have plowed into the barrier. It looks where it's going, more or less like you do.

Re: No way I could trust a self-driving car

[Scowler]

Taxi drivers send me to the wrong destination too... Should I trust them?

Re:No way I could trust a self-driving car

[tlhIngan]

Congratulations on only entering destinations that have existed for at least a year! Those of us living in "newer" portions of cities enjoy the constant joy of having our street "not exist" on Google Maps/Mapquest/wherever people are going now besides those.

That's really more the fault of Google than anyone else - Google's map updates are really slow since they started doing it themselves. When Google was paying Navteq or Tele-Atlas for the map data, it usually updates within 6 months (2 updates - they both do quarterly updates of map data)

I mean, Google still hasn't got my old house address in it. Something that was in every GPS update since 2010.

We mock Apple maps, but as long as Apple uses TomTom/TeleAtlas, at least the map data is updated fairly constantly.

Re:No way I could trust a self-driving car

[rasmusbr]

Let's just use OpenStreetMap. What could possibly go wrong?

Re:No way I could trust a self-driving car

[hey!]

New streets don't just appear overnight. They're planned, built, and then populated, a process that takes months. It is logistically possible, especially for wirelessly delivered services, for those streets to appear in real time as they're opened. Somebody has to pay for it, that's all. And the local highway departments have to work with the data providers of course.

If it's a year after a street opened and it's still not in Google maps, chances are there's something screwy with your local highway department releasing data -- maybe the county if you live in certain parts of the country.

Re:No way I could trust a self-driving car

[gweilo8888]

The problem is that even if new roads are reported to map vendors such as TeleAtlas, they take an interminable amount of time to add them.

I have a subscription to software using current TomTom maps. Before that, I had a dedicated TomTom device with a map subscription. Using their Map Insight feature, I reported a new subdivision to TeleAtlas -- which is *owned* by TomTom, remember. I provided them with complete details on the location of the new roads involved, or at least, as complete as is possible with their site.

Two years later, the roads still hadn't been added in my TomTom, at the point where I stopped my maps subscription. Courtesy of the third-party software using TomTom maps (Sygic), I can report that it was about 3.5 years from report until the roads actually got added to the map.

That huge lag -- even when the map vendor and the GPS device maker are essentially one and the same -- is why people have trouble with their GPS. That, and the fact that many of these maps are clearly made from satellite / aerial imagery, rather than from somebody actually pounding pavement -- and so when roads come close enough to each other beneath an obstruction (trees or whatever), the map maker believes the roads to be linked when they're not. (Or in more than one case where I live, their maps report somebody's private driveways between two roads as being a public road.)

And this, coupled with liability, is what's going to do in autonomous cars. 99.999% accurate and safe isn't enough when it comes to an autonomous car, and if you expect the driver to take over when the car gets it wrong, the driver might just as well be driving in the first place. As soon as the first reports come in of people dying because their autonomous cars autonomously crashed, the lawyers will have a field day at the auto makers' expense -- and the auto makers' lawyers know that too.

So they'll allow research and news coverage, because that makes the brand look futuristic and is effectively cheap advertising -- but you will not see these things on public roads in the hands of the public in your lifetime. You may see autonomous commercial vehicles -- especially those which drive predefined routes -- but not autonomous private cars. There's simply too much risk.

Re:No way I could trust a self-driving car

[JanneM]

Have you ever seen a stupid driver try to merge onto a freeway with their turn signal on and their foot on the brake? Or see three cars bumper-to-bumper trying to merge onto a busy freeway as a pack?

The beautiful thing is, self-driving cars will see this, due to their extensive sensor coverage. And they will have recordings available of the whole incident for later examination. It will be completely clear who was following all the regulations and exercizing judgement (the self-driving car) and who was driving irresponsibly or dangerously (the human driver).

Once self-driving cars hit the road in any numbers, it will become really, really expensive to try to be a jerk in traffic. No speeding. No lane cutting. No tailgating. No weaving. No nothing - just follow the traffic flow and the law to the letter and spirit or you will get reported. And of course your insurance rate will like go up, simply by being a human driver.

At which point there's of course little point in driving yourself any longer; you drive in exactly the same way as the self-driving cars, and you get there at the exact same time. But you have to sit there and drive, while the people around you are busy reading the morning news or throwing irate fowl at pigs.

Re:No way I could trust a self-driving car

[gweilo8888]

Great, it looks where it's going. Now, what happens when what you and I know as a road isn't recognized as one by the car? (Or vice versa, for that matter.)

Roads aren't standardized even from state to state or town to town, let alone country to country. Even in the USA, there are still many, many miles of roads that are completely unpaved, have no signage or markings at all, and are completely indistinguishable from somebody's driveway. Some of them even have foliage growing on them, or are so similar in surface to their surroundings that the only way to know where the road is is simply to pay attention to where the dirt is more packed.

You think your autonomous car will work in those conditions? I don't.

Re:No way I could trust a self-driving car

[Znork]

A self-driving car doesn't have to anticipate stupid drivers, it can keep appropriate distance and planning needed regardless of the inferred skill of the drivers. For example, why is anyone not exiting the freeway even in the lane that you know will get 30% bad mergers 500 meters ahead?

And frankly, yes, it's not hard to spot the stupid drivers and you could probably have algorithms for that. But the obvious stupid drivers aren't the danger, it's the good ones missing something as you won't be expecting them to do something idiotic. It's the biker you see slowing down and classifying as 'smart' who then rams straight ahead when you're looking in the other direction. It's the merger who perfectly accelerates up to the right speed, starts blinking, and then just doesn't see you. It's the guy who's stopped at the same red light a thousand times, but who's brain played a trick on him the 1001's time and had him think the green light for going ahead was the one that applied, not the red one for the left turn he just blew straight through.

A smart car will assume that everyone is stupid and that it, itself, has to be capable to counter any physically possible action. We should, as well, but the fact is that we're constantly hampered by our intelligence, assuming things that fit the general rules, anticipating based on experience, projecting states and emotions onto everything else. We shorten distances as we know that will make the driver ahead of us remember that he's forgotten to get out of the passing lane (or get uncomfortable enough to move out of the way). We maintain speeds that are usually ok, even in rain, when its dark, and our visibility is half of our stopping distance. If there's a stupid pedestrian about to pass we'll notice on the bump-bump. When we're the fourth car about to merge in to the freeway, do we stop and wait for the three idiots doing the bumper-to-bumper merge to get onto the freeway? No, most of us will at best give them a slightly wider berth, but most likely we'll see if we increase distance a bit and plan a higher acceleration and cut two lanes and get past them as fast as possible.

A well programmed smart car should know better than to get into a situation where an accident is inevitable, and it would at least theoretically be better at that. And the fact is, when the choice comes and it is inevitable I'd rather trust the car, because frankly, I've never, _ever_ been in that situation and I have absolutely no idea what I'd actually do or any time to think about it. The car, at least, might have knowledge of the statistics, its own capabilities and the physics of the situation.

Re:No way I could trust a self-driving car

[Anonymous Coward]

Yes, you're correct, very few humans "should" be licensed to drive based on competence.

Too bad, we built our society around having many drivers. So long as those drivers are humans, we have to relax or ignore that standard.

Hence the self-driving cars.

One thing nobody has mentioned throughout this is that drivers don't really have the capability to "drive" a modern car anyway. They're already fly-by-wire because no human wants four independent brake pedals (yes - the brakes on almost any modern car are all independent so as to make ABS, Stability Control, anti-roll and anti-drift work) for example. The self-driving car is just a continuation is a pre-existing trend. At first it will "help" you drive safely on long monotonous freeways, and then maybe it'll drop you off at the mall or airport and find a parking space. One day you'll realise that you haven't "driven" all day despite going two hours to a friend's house, and then you'll start wondering whether you should keep buying "manual" cars at all, especially since the insurance premium keeps going up...

Re:No way I could trust a self-driving car

[Anonymous Coward]

Some of what you've described isn't yet possible because of pre-existing expectations.

For example, normal drivers _expect_ to drive too close together to be safe. They _expect_ to drive too fast near pedestrians to react if the pedestrian unexpectedly steps out into the road. Autonomous vehicles can't refuse these expectations because drivers describe them as "too timid" and declare them faulty.

So AEB (Autonomous Emergency Braking) systems for example - the systems you'll see in most executive and more expensive family cars over the next few years - don't prevent all crashes at high speeds or in an urban environment. They could, but to do so they'd need to brake earlier and more aggressively, and to insist on a longer gap between vehicles, and the owners won't allow that. Because drivers insist on behaving dangerously, AEB systems will most often react AFTER a crash is inevitable, by slowing the vehicle and preparing for the impact. So a user's typical experience of AEB (if they experience it at all) will be that they thought everything was fine, doing maybe 80mph on open road, trying to get Billy to stop twisting his sister's arm, and then suddenly the car braked, the seatbelt pre-tensioner fired, and then they found themselves watching the airbag deflate and realising that the red Ford they'd be tailing was now somehow stopped and its rear end smashed up. Huh.

The AEB engineers also actually concluded that for overall safety it's best if the AEB brakes as late and as hard as will normally be safe, because if the emergency system just calmly prevents accidents without a big fuss people will just assume they're magically now brilliant drivers and can't get in a smash, which makes them take even more risks. So once the system decides to intervene it will always execute a full emergency stop with maximum braking to a halt, with the intention that the driver should be thinking "Fuck, that just saved my life" not "Oh it would have been fine anyway, I'm sure I would have slowed down enough".

Re:No way I could trust a self-driving car

[kyrsjo]

And even a really good human driver has bad days, get fatigued, distracted etc. - that's problems avoided with a computer.

Actually, even if this ends up being mostly highway-only, that covers the most important part. Highway driving is pretty boring anyway, and plagued by the problem that nothing happens for hours, until you suddenly hit the back of a queue or someone around you does something stupid, and you have to react fast. Doing that automatically would combine the convenience and privacy of a car with the comfort of a train. But it would have to be good enough to be trusted *completely* while on the highway, with at least one minutes notice before the driver has to take the wheel manually, in order to put the seat back into driving position and put away books / laptop / food etc.

Re:No way I could trust a self-driving car

[emj]

Once self-driving cars hit the road in any numbers, it will become really, really expensive to try to be a jerk in traffic. No speeding. No lane cutting. No tailgating. No weaving. No nothing - just follow the traffic flow and the law to the letter and spirit or you will get reported. And of course your insurance rate will like go up, simply by being a human driver.

It's not clear that they will be allowed to record all that, at least not in Sweden, which this story is about. Privacy problems and all that.

Re:No way I could trust a self-driving car

[AmiMoJo]

You know, these cars still have a steering wheel and pedals. There is a manual mode available.

Re:No way I could trust a self-driving car

[Znork]

You're right, of course, the assistive technology will be handicapped by driver expectations. Personally I expect the full blown automation to be less subject to that as it reverses the pattern of interference. The reluctance to interfere with a driver unless you're certain he's missing something goes both ways, and 'passengers' in a fully automated vehicle won't be more prone to back seat driving than anyone usually is in a taxi or a bus or other vehicle they're not themselves in control of.

Re:No way I could trust a self-driving car

[Anonymous Coward]

Actually, even if this ends up being mostly highway-only

It won't.

The article doesn't mention it, but here is another link about the project. It is essentially the old Google-tech that has been tested on highways but improved to handle high traffic situations in inner cities.

Re:No way I could trust a self-driving car

[Reziac]

And how to have 'fun' with self-driving cars:

Crowd 'em.

Right into the ditch.

It should be fairly easy to do, since the SDC won't have the ability to be 'rude' and push ahead or sneak into a narrow gap to get away from the crowder.

Having had to hit the ditch a few times to avoid a head-on, I appreciate the necessity of drivers being able to make judgment calls.

And I wonder how much this is a solution in search of a problem. A while back someone here piped up with the "accidents per mile driven" stats, and the risk factor was ridiculously low, considering how many millions of miles we drive in the U.S.

Re:No way I could trust a self-driving car

[Reziac]

And then there's the issue of dedicated streets that don't actually exist. A fun example in Los Angeles runs right through the middle of the big Water & Power facility.

And there's the ongoing example of "Tail of the Dragon" in Tennessee/North Carolina... a perfectly valid road, a U.S. Route no less, yet not suitable for a significant fraction of traffic sent via that 'shortcut' by their GPS.

http://tailofthedragon.com/dra...

Re:No way I could trust a self-driving car

[Reziac]

How quickly can one switch to manual mode? For anything that's truly an emergency, you may have bare seconds to respond, or perhaps only a fraction of a second. Not enough time to flick a switch and move your hands and feet to the controls, but enough to react if you're already "in position".

only viable in small countries

[Anonymous Coward]

You need to augment the roads for this tech to really work.

Here in the southern US, we can't even fill pot holes. The roads and bridges are crumbling beneath us.

Not to mention that there is a lot of miles of road in total that would need to be updated to work with this stuff.

Basically, for the same reason we have shitty internet, we're not going to have this stuff by 2020.

Re:only viable in small countries

[Spy Handler]

autonomous cars don't suffer any more damage from potholes than regular cars.

Re:only viable in small countries

[Anonymous Coward]

I bet they will hit more of them than a human.

So yeah, they do.

Re:only viable in small countries

[Anonymous Coward]

You mean the computer with cameras facing everywhere that is always paying 100% attention, will hit more potholes than a half-awake driver looking wherever the fuck he feels like at any given moment?

Re:only viable in small countries

[Anonymous Coward]

How does radar detect negative space?[like pot holes]

Is the car aware of where each tire is relative to a pothole, and can it plot a course around it?

Things like obstructions, other cars are easy-ish[What about debris. I know I can safely run over a plastic bag, but will the car freakout if one blows across the road?]. But missing pieces of road are another thing all together.

Re:only viable in small countries

[CanHasDIY]

You mean the computer with cameras facing everywhere that is always paying 100% attention, will hit more potholes than a half-awake driver looking wherever the fuck he feels like at any given moment?

No, I'm pretty sure he means that the computer is going to put the "stay between the lane lines" rule before the "dodge the pothole" rule in the list.

But please, don't let a little thing like reason keep you from being an unbearable dickhead about it, that will surely convince the fence-sitters to side with you.

Ass.

Re:only viable in small countries

[Anonymous Coward]

What happened when a auto car missreads a child as safe to drive over?

Re:only viable in small countries

[Ralph Wiggam]

You need to augment the roads for this tech to really work.

Citation?

Re:only viable in small countries

[Anonymous Coward]

Actually, what is far more likely is that California will mandate that self-driving technology is implemented so that potholes can be avoided. Now, suddenly, the government has even less incentive to spend our collected tax dollars on potholes. That would involve money flowing in the direction of the elite powerful people into a project that benefits the average citizen.
Instead, if a law is made that helps money go into the opposite direction, then corporations benefit as people are coerced into spending more money on some fancy new feature that proponents say will make everybody's lives better.

And, regarding the other comment that cars driven by AIs will hit more potholes than cars driven by humans: that may be true, for the first version. That is, until someone comes up with something that is theoretically better, and politically pleasant enough that California makes a new law demanding that all manufacturers start supporting an upgraded standard that has so many concessions to certain intellectual property that nobody can make a compliant version using only free software.

Re:only viable in small countries

[Anonymous Coward]

Source: Common sense and a computing background.

The radar's need either very clear/clean lines or electronic markers to know where they are. Ideally both.
Most roads, at least around here, are often covered by dirt, oil and debris, have very faded lane markers, or missing them entirely.

Re:only viable in small countries

[wonkey_monkey]

How does radar detect negative space?[like pot holes]

It doesn't. It detects that the distance to the ground is greater than the surrounding road and infers the "negative space" (whatever that means) from that information.

Is the car aware of where each tire is relative to a pothole, and can it plot a course around it?

Why wouldn't it be?

Re:only viable in small countries

[wonkey_monkey]

No, I'm pretty sure he means that the computer is going to put the "stay between the lane lines" rule before the "dodge the pothole" rule in the list.

There you go, you've solved the problem. Put the "dodge the pothole" rule above the "stay between the lane lines" rule, but below the "mow down little timmy" rule.

Re:only viable in small countries

[Ralph Wiggam]

I'm not sure why you think your computing background means you know dick about autonomous cars.

They don't use radar at all and lane lines are only one of multiple visual cues used.

Re: only viable in small countries

[Anonymous Coward]

The auto car will certainly kill fewer kids than a human driver, most likely engrossed in texting someone.

Re:only viable in small countries

[Anonymous Coward]

https://en.wikipedia.org/wiki/Google_driverless_car

The first line in the Technology section....laser radar. Two seconds of Google search illuminates who doesn't know dick.

Re:only viable in small countries

[Ralph Wiggam]

Apparently neither you nor the guy who wrote that wikipedia page know that lidar is completely different from radar. The "ra" stands for radio.

Re: only viable in small countries

[Anonymous Coward]

Computing background doesn't necessarily provide much insight in these situations. More of a DSP area

Re:only viable in small countries

[tolydude]

There you go, you've solved the problem. Put the "dodge the pothole" rule above the "stay between the lane lines" rule, but below the "mow down little timmy" rule.

Sucks to be little timmy.

Re:only viable in small countries

[gweilo8888]

Not to mention that you quite likely know where many potholes are locally, and you avoid them without even thinking about it, whether you can actually see them or not. Your autonomous car can't tell a puddle from a puddle with a pothole beneath it. Do you want to drive in a car that has to swerve around puddles, just in case? Because if not, you're going to be hitting more potholes, like it or not.

Re:only viable in small countries

[kyrsjo]

Unless it can check a "pothole database" in the cloud, which will tell it where it needs to watch out.

Re:only viable in small countries

[Reziac]

And potholes do wonders for your mechanic's bank balance.

I challenge a SDC to make it down 10th Avenue South in Great Falls MT after a winter's freeze/thaw cycles have turned this main-drag (for both cars and trucks) into an obstacle course (which happens despite intensive maintenance). We used to not-quite-joke that there were VW Bugs swimming in the larger potholes.

highly autonomous mode ?

[Anonymous Coward]

wtf does "highly autonomous mode" mean ?
Does that mean I still have to be there and pay attention ? Then it is useless. Anything but 100% autonomous is useless. If I can't sleep in the back of the car while the car drives, then what is the point ?

Re:highly autonomous mode ?

[greenwow]

It's a Republican-created term to make these dangerous things sound even more sinister. They hate technology so they hate these cars. Of course, they also hate these cars because it makes driving safer. Their kind loves dangerous things like fast cars and guns. They want to force those things down the throats of the rest of the population. Fortunately, most people in the US are sensible and not members of that awful group, but it doesn't stop them from ruling over us since they're the ones with the money.

Re:highly autonomous mode ?

[scorp1us]

It's the mode created by lawyers to maintain that you are responsible for the vehicle's operation. because if the software fails, it's Volvo's fault. Imagine you drive through a oil slick, the vehicle takes a corner too fast before becoming aware of the reduced traction. In automatic mode, it's the car's fault, in "highly autonomous mode" it's still yours.

I'm driving!

[Anonymous Coward]

You can have my car keys when you pry them from my cold dead hands.

Re:I'm driving!

[ArcadeMan]

Good news everyone! Cars won't need keys in the future!

Re:I'm driving!

[canadiannomad]

You can have my car keys when you pry them from my cold dead hands.

Next on YouTube:
"The autonomous car's video, audio and 3D representations of exactly what happened as the fire crews pulled the keys from AC's cold dead hands, following an accident caused by human error. The passengers of the autonomous car all escaped unscathed."

Re:I'm driving!

[FlyHelicopters]

The future? Heck, my truck doesn't use keys today! I just walk up, put my hand in the handle and press the button on the door, it unlocks. The seat knows who I am and moves to my position, then I just press the button on the panel and the truck turns on, then I go drive.

When I stop and get out, the truck autolocks about 30 seconds after the last person leaves, it will even warn me if I leave the little key fob thing in the truck by flashing the lights and honking the horn.

I can also turn it on and off using my phone and I can remote start and stop it using my phone, about the only thing the phone won't do is drive it via remote control.

Re:I'm driving!

[CanHasDIY]

The future? Heck, my truck doesn't use keys today! I just walk up, put my hand in the handle and press the button on the door, it unlocks. The seat knows who I am and moves to my position, then I just press the button on the panel and the truck turns on, then I go drive.

When I stop and get out, the truck autolocks about 30 seconds after the last person leaves, it will even warn me if I leave the little key fob thing in the truck by flashing the lights and honking the horn.

I can also turn it on and off using my phone and I can remote start and stop it using my phone, about the only thing the phone won't do is drive it via remote control.

The downside, as my brother-in-law found out the hard way, is that if the battery in your keyless fob starts to die when you're driving down the road, it also shuts the engine down.

At least, it does on a Mazda MX-9 (or whatever their big-ass SUV is). YMMV. I'll stick to physical contact authentication methods, thanks.

Re:I'm driving!

[Anonymous Coward]

I don't even trust power locks and windows to work reliably enough. I can't imagine having the fate of my ignition in the hands of a CR2016 watch battery.

[Or GM's springs.]

Re:I'm driving!

[FlyHelicopters]

That does not happen, or it shouldn't anyway...

It will alert you that it can't detect the key, but the engine isn't going to shut off.

If you take the key inside when the truck is running, it will ding at you every minute and if you try and turn the truck off, it will warn you that a restart will require the key.

If the battery dies in the fob, there is a small physical key inside that can be used to get in the truck if required.

Keep in mind also that OnStar can remotely start the truck for you as well if you have a problem, and you can drive it that way without the key fob if you know the password.

Just get in (unlock it either by using your phone, or calling OnStar, then press the blue button and tell the adviser that your key fob isn't working, they'll ask you your security question then turn the truck on for you.

It really isn't as hard as many people make it out to be.

Re:I'm driving!

[FlyHelicopters]

Well then, good luck buying anything new these days... You'll be hard pressed to find a car or truck that doesn't come with power windows and door locks these days...

As a side note, besides being able to open the truck using a phone (either via app or calling OnStar), you can open it using the hidden physical key in the keyless fob (they all have one as a backup)...

On top of that... you can find a CR2016 battery... well, just about everywhere... So getting a replacement is not exactly hard...

Let's hope it goes better than last time.

[scorp1us]

Last time

Someone might die

[Anonymous Coward]

There could be an accident where someone dies so I'll stick to having my mom drive me to school.

Lane Following

[tompaulco]

How well does lane following work when the road is under construction and the lane shifts over. The lines demarking the lane don't necessarily shift. What about road hazards such as a sinkhole has destroyed half of the lane?
Just like software, these seemed designed for perfect conditions, and perfect conditions only exist in controlled environments, and sometimes not even then.

Re:Lane Following

[greenwow]

Or here in Seattle where the city is too cheap to actually put lines on most streets. Instead, they glue rough lumps of clay to the road to make it more dangerous to ride a motorcycle, scooter, or bicycle. The people that rule this city hate those of us that ride smaller more fuel-efficient vehicles so they try to kill us. Not only are they designed to make it impossible to get traction on a two-wheel vehicle, they are also designed to become projectiles when they come loose from the road surface. I've lost two face shields from getting hit by those Republican bike killers. There's no way an autonomous car can follow those glued lumps of clay.

Re:Lane Following

[Anonymous Coward]

What about snow?

What we really need

[Anonymous Coward]

is stricter licensing requirements and periodic re-testing. Also, actual criminal fines[like the good old days] for speeding, redlight running, etc. We need to actually inspect vehicles at annual[now increasingly every two years] inspections.

We need to encourage schools and corporations to allow telecommuting and 24 hour shifts to reduce / spread out the load on the road at any given time.

We need to encourage living closer to work/school/etc. When this happens mass transit becomes more effective.

But first, we need to start thinking more about each other than ourselves. We need to vote in verifiably non-sociopathic representatives and bankrupt the Walmarts/Comcasts/etc of the world.

Re:What we really need

[Mr D from 63]

If they can get this car passed by my old Driver Ed teacher, Mrs. Pinnick, then its ready to go!

But seriously, there are so many conditions and variables out there that even the most extensive testing cannot cover all scenarios. And who will be the first company to officially deploy this technology and risk the lawsuits from the first death/mishap.

And if I have to stay behind that obnoxious bike rider for 15 miles because the car is too safe to pass, then I just might.... say something unpleasant.

Re:What we really need

[Dare nMc]

>deploy this technology and risk the lawsuits from the first death/mishap.

Insurance can take up the banner. Google shows the car is safer, then requires proper insurance that covers them. If the statistics pan out, insurance companies will charge less, despite the risk of a bigger pay-out from a jury on a single case.

Re:What we really need

[KingOfBLASH]

OK but this really becomes a question of the car "failing safely." If the car is able to judge when it is unable to drive safely (say, a blizzard) and can then stop and tell the human driver to take over, it'll be OK. If the car gets confused and just shuts down on a highway where everyone is going 80MPH, the car is going to create a dangerous situation when it's unable to continue. (Although, in all fairness, breakdowns in conventional cars do occur on the highway and it doesn't necessarily mean death to the driver)

The big problem, as I see it, is that the reason I want an autonomous car is because I want to go sit in the back seat, light up a cigar, pour myself a scotch, and read a good book on a long road trip. However, if you still need a driver capable of taking over and driving, it kind of kills it. And I would bet it will be much more annoying to be watching a car drive itself and think "OK do I have to take over now" then just driving myself.

SyFy movie?

[slapout]

Why do I feel like I'm watching the start of a bad SyFy movie?

Re:SyFy movie?

[ebno-10db]

Why do you say it's bad? I think it's going to be a good flick (as long as you stay off the roads).

This is a bad idea

[edxwelch]

Aren't they worried that they might become sentient and turn on their masters?
Kill all humans!

Blah blah blah

[Anonymous Coward]

People have been predicting my immanent death behind the wheel (or the handlebars of my motorbike) for well over two decades now. You can be just as wrong as them now.

Re:Blah blah blah

[Anonymous Coward]

Don't confuse chance with control. You are a lot harder to see on a motorcycle even if you are a responsible rider. The physics is going to almost always be on the side of your opponent when a crash happens too.

Re:Blah blah blah

[wonkey_monkey]

You can be just as wrong as them for now.

FTFY.

Why luxury safer electric cars should be free

[Paul Fernhout]

https://groups.google.com/d/ms...
"This essay explain why luxury safer electric (or plug-in hybrid) cars should be free-to-the-user at the point of sale in the USA, and why this will reduce US taxes overall. Essentially, unsafe gasoline-powered automobiles in the USA pose a high cost on society (accidents, injuries, pollution, defense), and the costs of making better cars would pay for themselves and then some. This essay is an example of using post-scarcity ideology to understand the scarcity-oriented ideological assumptions in our society and how those outdated scarcity assumptions are costing our society in terms of creating and maintaining artificial scarcity."

uh no

[WindBourne]

Mercedes JUST did their first back-walk at self-driving cars in sept, 2013.

who actually wants these?

[cascadingstylesheet]

Just curious, is there really much demand for self-driving cars?

Yes, I know who wants to build them, but who (as other than a tech toy) really wants to buy them and, er, not drive them?