By on March 18, 2014

Jurvetson_Google_driverless_car_trimmed

TTAC reader and former auto journalist Michael Banovsky writes about the inexorable move towards autonomous cars

Autonomous cars are already here.

It doesn’t matter if you’re testing an actual Google Car or cruising the Keys in a Pagoda-roof 230 SL, CUVing the kids to Hot Yoga or signing “11” on a deserted road. Autonomous cars are here, the debate is done, so enjoy driving while you still can.

Let’s start with a story.

I was driving to work and glanced in my rearview and noticed a lady talking on a cell phone. Is that a chil…yes, that’s a child-in-child-seat, too.

We were at a moderate speed, we stopped, we got going again…and she didn’t hit me. I even watched, two minutes later, as she put the phone down and resumed the school run.

What was I supposed to do, publicly shame her? Call the cops, telling them someone was making a call—a possibly important one—and they should speed over, tout de suite?

This happens all the time, of course, all over the world. Are we to vilify everyone who safely makes a call or text while behind the wheel? Drives drunk? Drives high? Drinks coffee without spilling it? Changes the radio station without crashing?

Speeds?

I don’t think so. That would be—caution, nasty word – surveillance, and we’re probably going to give up driving before it’s monitored or taken away, anyway.

Here’s why: Any anti-social and anti-public safety behaviours* are drivers showing they’ve chosen something else over operating a vehicle. Taking a call while driving is proof, proven thousands of times a second, that we feel talking on a phone is as important to us as driving.

For a driverless future to happen, two things need to happen. First, non-compliance with road laws and rising costs will make driving much more expensive—to say nothing of fuel prices. Second, technology will make it possible.

Now tell me either is unlikely.

The key to adopting driverless cars without outcry is to make drivers feel like they have a choice. The lady I saw talking on her phone? If you could have given her a big green “Autonomous” button, I bet she’d have pushed it before taking that call.

Fines for not complying will keep increasing, making a driverless car system—either built-in or aftermarket— seem cheap in comparison. The aftermarket devices will become so small as to be unnoticeable. What will stop companies from offering ad-supported ones? “Saving $20 on groceries this week will only take 9 minutes, Ms. Greer. Would you like me to set a route?”

Autonomous vehicles could allow us to:

  • – Safely accept phone calls
  • – Safely interact with passengers
  • – Safely navigate through stressful or dangerous driving conditions
  • – Appoint an adult bus monitor instead of driver, making the now-autonomous school bus safer
  • – Drive your drunk ass home
  • – Travel more quickly on highways (what government would argue against higher speeds if they were sure crashing wasn’t possible. Yes, your car will drive faster than you.)
  • – Substantially reduce insurance premiums
  • – Substantially improve pedestrian and cyclist safety
  • – Substantially improve fleet-wide fuel economy
  • – Revolutionize semi-public transit, like airport shuttles and taxis
  • – Send our vehicles for service while we’re at work
  • – Offer incentives to shop in certain stores, or drive in certain places
  • …and many, many other things.

Roads were humanity’s last great analog system, until of course we started mapping things digitally. GPS and Google Streetview for our system of roads. Radar, specialized cameras, sensors for vehicles themselves. The vehicles are irrelevant—at the point machines move for themselves, does it matter if it’s a cement truck or smart fortwo? Does it matter if the data required to move a machine comes from a satellite or the car in front?

Once machines can read the road surface, signs, and conditions accurately (and reliably), these systems will flourish, and the vast majority of motorists will benefit.

Don’t like it? Don’t speed. Don’t use your cell phone. Drive more smoothly. Don’t crash. And tell millions of others the same. Then keep it up for the foreseeable future.

A future where driver-less cars outnumber driver-with cars isn’t crazy. It’s certainty, certainly if drivers keep breaking the rules. Statistics proving how bad we are at driving will allow the technology a foothold, and a few machine generations will work out most problems.

Advertising will take care of the rest.

What, did you think for a moment that companies would allow one of our last, great freedoms—driving—to remain free from monetization forever? “Driving” will become “moving people around.”

If you’re in doubt, take a few minutes and read US Patent #8630897. Search for “Autonomous.”

*As defined by our road laws—if you don’t like them, change them! (Ha.)

Get the latest TTAC e-Newsletter!

145 Comments on “Ur-Turn: Autonomous Cars Are Already Here...”


  • avatar
    raph

    Bah… I live for breaking the rules. It keeps me from doing something really dangerous!

    Probably a big reason why I view autonomous cars as anathema and only peripherally understand driving solely as an inconvenience and chore.

  • avatar
    imag

    Fines aren’t even necessary. Most of the stuff on your list isn’t necessary.

    As soon as my wife sees a friend get out of a self-driving car, she will want one. My guess is that she is not alone. Most people don’t like driving. Heck, I don’t even really like driving unless I am out on an empty highway or on the track.

    They will catch on because they are cool. Manufacturers will get them in the hands of celebrities and people will want them. They will try them for ten minutes and never want to go back.

    After all – who wouldn’t want a personal chauffeur – without the guilt or the cost?

    • 0 avatar

      When consumer desire shifts in such a way that people realize they won’t get to work faster in a sports car, autonomous vehicles and electrification are the perfect marriage of technology. Better vehicle packaging, smart (ubiquitous) charging, completely quiet running.

      • 0 avatar
        baconator

        +1 on this. I love cars, love driving, but my next car will likely be whatever lets me drive in the carpool lane. A plug-in Prius reaches about twice the top speed of a Ferrari 458 if they’re both in rush-hour traffic on the 101 freeway. And if I could buy an autonomous car and get that hour of my life back…huzzah.

      • 0 avatar
        imag

        In a way, it will be a lot like what we saw with cell phones. They initially were justified based upon making phone calls away from the office or home.

        Now, though, the whole idea of being tied to a landline is ridiculous. The selling proposition has morphed into apps, videos, personal computing tasks, instant access to online information and accounts. It will likely evolve into always-on computing and life recording.

        Likewise, the selling proposition of a self driving car will be less about whether or not it can drive itself, and more about the total configuration change of the automobile into an autonomous mobile space. As they become ubiquitous, you can eliminate traffic signals, road signs, and some of the traditional road and parking geography. Leasing and contracts become the standard upgrade mechanism, and the car features are less about driving and more about comfort and mobile entertainment.

        • 0 avatar

          +1

        • 0 avatar

          @imag
          >>>As they become ubiquitous, you can eliminate traffic signals, road signs, and some of the traditional road and parking geography.

          I think it may be more complicated than that, at least depending on how the autonomous tech works. If the cars have to communicate with each other, pedestrians and bicycles may become the weak link in that system, forcing the continuation of traffic signals.

          Also, if autonomous cars are not foolproof, or pretty damn close, people are going to have to be equipped to take over, which means learning to drive well, which means, well, doing a lot of driving.

          While I would not bet heavily against autonomous cars, I will believe it when I see it.

          • 0 avatar
            imag

            I think it will be a long time before traffic signal elimination. Self drivers are being developed without the need for communication or special signals or lanes. And if you read about what the Google car can see, pedestrians and bikes are actually much safer around an autonomous vehicle than a human-driven one. I do think there will be other city infrastructure changes, like pick-up and drop-off zones around buildings.

            I also think it will be a bit of time before we lose the steering wheel, but I would bet that it will happen by 2040.

            To avoid lawsuits, the self driving cars will have to be far better than human drivers before they are implemented. Once the safety level is accepted, human drivers might actually look like the weak link, and be *discouraged* from driving. 34K people got killed last year in motor vehicle accidents last year. Autonomous vehicles pay attention all the time; they can monitor for things like low tire pressure and mechanical failures; and they would drive cautiously at all times. It could look downright selfish to drive your own.

    • 0 avatar
      CoreyDL

      Enter a MASSIVE market for LWB cars.

  • avatar

    So sad. The world will never be the same and it won’t be the world I grew up in. Feeling really old as i’m sure the young ones will be all over this.

    • 0 avatar
      wmba

      +1.

      Why doesn’t Google invent something useful? Like an autonomous car that cleans off 8 inches of snow in the morning, has little robot arms that come out of the bodywork to clean off the sensors so that the various cameras and radar can work reliably, deploys a snowplow to clean the driveway, and has a reliable pot hole avoider that prevents head on crashes as it weaves around urban decay, etc. etc.

      People would pay money for these features, but the folks out in California don’t see much snow in everyday driving, and would rather chase ephemera.

      This whole autonomous car thing has not been thought through, in my opinion. If people embrace the idea, they’re going to really be upset at having to provide ANY input other than to flounce out of their front door and be magically whisked to work.

  • avatar
    imag

    The interesting issue, to me, is how things change with self driving cars.

    I actually started a thread on the forum about it here, for those interested: http://www.thetruthaboutcars.com/forum/new-future-product-and-industry-discussion/the-truth-about-self-driving-cars/

    • 0 avatar

      You’ve nailed it. I love driving, but see that things will change. For better? Probably. For worse? Probably.

      Also, if crashing isn’t a big part of driving any more, maybe UFC and Indycar will merge to become some sort of arena or old racetrack bloodsport…

    • 0 avatar

      BTW, I posted on your thread before this article but for whatever reason it’s still “awaiting moderation”, so here it is:

      > City drivers will ditch their cars, but not suburbanites. Uber won’t take over everything. People will still want their own little living room, with their own stuff in it.

      To the contrary once cars truly self-drive w/o caretaker automatic chauffeur services will be common simply due to cost. Why take the hit on a $30k+ depreciating asset alone when it can come shortly at beck and call? A modular vehicle with a passenger compartment left at your own place can serve for those who need their own living space.

      • 0 avatar
        imag

        Yeah – I was wondering about your comment. I thought maybe you swore at me or something ;)

        I agree that auto chauffeur systems will take off in many markets, but I think many people, especially commuters, will still want their own vehicles. People like to keep stuff in their cars, and their space itself is a kind of status symbol.

        Also, the chauffeur system depends on utilization of the car by multiple people on the same day. It breaks down when a vast percentage of the cars get used twice per day at exactly the same times. There simply isn’t a need for as many cars at noon as there is at 8 pm and 5 pm.

        I do think that leasing/contracting your vehicle will be a standard thing due to the depreciation, just like what we see with phones. The older vehicles will get moved to less-wealthy areas or countries.

        • 0 avatar

          > It breaks down when a vast percentage of the cars get used twice per day at exactly the same times. There simply isn’t a need for as many cars at noon as there is at 8 pm and 5 pm.

          Couple of mitigating factors:

          1. Rides will be priced by demand. Pch101 in the last discussion on this make the behavior econ case that making costs explicit tends to drive (irrational) decisions for better or worse.

          http://www.thetruthaboutcars.com/2014/02/analysts-peak-car-to-arrive-by-2020s/#comment-2867873

          2. Such a system trivially transitions between chauffeur and ride-share services. Get to work with others and split the costs. This is a separate issue though since something like Lyft can presumably enable this well before robot cars.

          In the end people with the means will always buy the car, but technology can provide a cheaper and just as convenient alternative.

          • 0 avatar
            imag

            I agree with PCH; people irrationally favor certain approaches to pricing.

            For one thing, people generally prefer fixed price models over variable rates.

            Public transportation has the commute problem now. They don’t try to change people’s commutes by charging differential rates. They just run a bunch of trains into cities and then run them out again at the end of the day. Taxis deal with varied demand as well. People would be cry bloody murder if they charged based on demand.

            Much of the time, people also favor fixed price “all you can eat” models over variable cost-per-use models. Look at Netflix versus Pay Per View. Occasionally, lump sum advance payment models work for people, but no one has been able to get micropayments to work online.

            I think that some reasonably significant percentage of folks are going to want to “own” (even through a lease) their car. They spend a lot of time in it, and they are going to want to feel like it is their space. I agree that it is irrational ;)

  • avatar
    KixStart

    Self-driving cars will liberate the elderly.

    And there’s going to be more and more elderly, so self-driving cars are inevitable.

    • 0 avatar
      imag

      And:
      – the blind
      – the disabled
      – those with temporary disabilities

      • 0 avatar
        CoreyDL

        -drunks
        -children

        I can just see it now. “Twelve year old boy travels 350 miles in car while parents at work to visit Disney World, alone.”

        • 0 avatar

          Not if “to drive” requires some bio-authentication, and little Jonny hasn’t yet figured out how to scoop eyeballs out of his parents.

          • 0 avatar
            KixStart

            Bio-authentication isn’t necessary, we’ll still have keys of some sort.

            Sufficiently motivated twelve year old kids already sometimes make unauthorized use of motor vehicles.

            I would expect the autonomous car would report its position to the owner on demand. In fact, it would be helpful to be able to send the car on errands by itself.

            “Car! Fetch Jimmy!”

    • 0 avatar
      GoesLikeStink

      Think of the children. No more need of the Mom taxi. “Car, go get Timmy from school and take him to the game. Then go get Dad from work while I finish this corporate merger”

  • avatar
    Landcrusher

    You could have written this article in the nineties. Not much would be different. Sorry, but I was looking for something new to read about.

    NS

  • avatar

    I agree that the general movement of society is strongly in the direction of autonomous cars vs non-autonomous cars. You can already see the potential with divided lanes between HOV/HOT lanes vs. regular traffic lanes. My guess is that autonomous cars will show up in these divided, separate lanes first. Of course that would mean we would eventually have to end up with something like this for non-autonomous cars:

    http://www.theonion.com/articles/dot-creates-new-lane-for-reckless-drivers,2405/

    • 0 avatar
      PandaBear

      Indeed.

      The biggest problem with autonomous car is how will it deal with worst case scenario (safety). Will it stop? Slow down? How much safety margin will build in to make sure that if it can’t “see” the road or obstacle well it will slow down and stop? How to deal with false positive, etc? All of the corner cases.

      So for it to be truly useful, first the road needs to be reworked to make it idiot proof for these cars, that means taking out lane. Sharing lane between autonomous car and non autonomous means idiot mistakes and finger pointing (to computer).

      I think autonomous car will first come on long haul trucking with dedicated road or big blinking industrial light you see on manufacturing lines, from truck stops to truck stops. The driver will be needed to park it, move it around town, and can sleep in between (if not responding before the auto lane end the truck will park on the side of the road waiting for response). This means a driver can “drive” 20 hours a day while sleeping 12 hours inside the truck and still be refreshed enough to be safe, and still make money when the companies will have reduced cost.

  • avatar
    DC Bruce

    Two thumbs up for Banovsky! One thumb for writing an interesting article about this topic from a perspective I haven’t seen before and a second thumb for eliciting some interesting and useful comments — especially KixStart’s. Those who have elderly parents frequently have stories to tell about the angst associated with getting Dad or Mom to give up driving — because he/she is a hazard to others. It’s not that most old folks think driving is fun; they don’t. It’s the loss of mobility that pains them, and rightly so.

    For the vast majority of people who own them, a car is transportation. It’s not necessarily fun; it’s not necessarily a signal of your personal wealth (or lack thereof). So, if autonomous cars make “transportation” better (i.e. cheaper, faster, safer, easier), then that’s a good thing for them. Frankly, in all but a handful of super-dense cities, autonomous cars would be a far better public investment that 19th-century style mass transit rail systems, which are inflexible and expensive to build and operate. Right now, the only advantage these systems have is that they’re faster than busses because they don’t have to deal with street traffic. Improve the speed and orderliness of street traffic, and buses look like the better alternative even in fairly high density cities.

    In suburbanized areas, autonomous cars would be the best. And, yes, in that service, electrics would be best, too.

    • 0 avatar
      dal20402

      “Improve the speed and orderliness of street traffic, and buses look like the better alternative even in fairly high density cities.”

      The problem is that in any city denser than an ordinary suburb there is just not enough space for every person who is traveling from A to B to be surrounded by a 15x7x6 box of metal. Even with autonomous cars that perfectly optimize traffic flow, you will still need both mass transit and less space-intensive modes such as walking and cycling.

      The market will enforce this need through 1) congestion and 2) high prices to either park the autonomous car or to get it out of the city to a place where it can park.

      • 0 avatar
        bumpy ii

        Which is why you split the difference with autonomous minibuses.

        • 0 avatar
          dal20402

          They solve the parking issue, but only do a bit for the congestion issue. We will still need a grade-separated people-moving network in major cities. (But the trains or BRT buses, like everything else, can be driverless!)

        • 0 avatar
          GoesLikeStink

          Car sharing could become a lot more popular especially with people who have complementary schedules. Why should a car sit in a garage all day if it can go give someone else a ride?

          • 0 avatar
            KixStart

            I’m liking this idea.

            Call up and request a trip in an autonomous taxi. Pay less if you’re willing to share.

            There will be an app for that.

          • 0 avatar
            CoreyDL

            New for 2035, the BMW 4-Series Wagon, with Autono+.

            For just $7,895 BMW will install Autono+, which gives you the right of way in 45% more traffic situations!

            To obtain right of way in 74% of situations, Autono+1 is available in all 6-Series GT x-Drives, and 725 FWD models.

      • 0 avatar
        GoesLikeStink

        Traffic happens because everyone is on the road at the same time. I do not see how the same number or more cars is going to get everyone to work any quicker. It will probably be slower. If the infrastructure does not change and all of the robo cars are doing the speed limit, make full stops at all signs and waits for safer openings on turns, traffic will move slower than it does now.

        • 0 avatar
          JuniperBug

          You underestimate to what degree human drivers fail at optimizing traffic flow. A computerized car which can communicate with the cars around it can do a much better job of optimizing flow. Using the stop sign example: upon approaching an intersection, two cars can already communicate with each other, determine which one will go first, and at least one of them won’t have to stop at all. Computerized cars won’t sit in the left lane holding up traffic behind them, and will be able to merge efficiently onto highways without anyone having to slam the brakes.

          In theory, some of this could be done by competent human drivers, but at least in North America, thinking about the good of the collective traffic flow largely eludes us. Even assuming human drivers who know what they’re doing, they can’t communicate and synchronize with traffic control devices and other cars approaching nearby the way computerized cars could.

          • 0 avatar
            GoesLikeStink

            “Using the stop sign example: upon approaching an intersection, two cars can already communicate with each other, determine which one will go first, and at least one of them won’t have to stop at all.” Killing children and other pedestians left and right.
            Also for this to work all cars would have to be on this network. That aint gonna happen. We cant even get the pre smog cars off the streets. We are not going to get 100% compliance and that means all the robo cars will have to obey the current laws and avoid the humans who drive like crap by driving even more cautously and slowly.

        • 0 avatar

          Possibly. That said, who needs stop signs when the car can see around the corner? Or there’s some sort of proximity sensor that alerts a car to another, preventing any accident?

          You could conceivably navigate as quickly as the computer and conditions would permit. Maybe pay a little extra to use a priority lane. Or pay a little extra to have autonomous vehicles in front move over…

          In traffic, to make a meaningful difference all you’d need are four or five autonomous cars, spread out across all lanes to “pace” the rest in order to eliminate the accordion effect.

          • 0 avatar
            GoesLikeStink

            “That said, who needs stop signs when the car can see around the corner? Or there’s some sort of proximity sensor that alerts a car to another, preventing any accident?” You do not live in a residential neighborhood or have children or pets I guess. There are a lot more things to worry about on the road than other cars. I can see this being a good technology on the much more controlled enviroment of the freeway, maybe making them useful again. Which would in turn free up surface streets. Imagine a 405 that was not a parking lot.

        • 0 avatar
          jmo

          They’ve done studies and if as few as %15 of people were in autonomous cars traffic would fall dramatically. It turns out that the vast majority of traffic is caused by pressure waves that develop. For a very simple example – think of a traffic light turning green. Do all 10 cars suddenly begin to move forward at the same time? No, car 1 moves, then car 2 etc.

          On the highway, a disturbance occurs and people brake excessively and then when the obstacle clears then accelerate one at a time. With as few a 15% of cars autonomous, the computers would damp these pressure waves and dramatically reduce traffic.

          http://web.mit.edu/newsoffice/2009/traffic-0609.html

      • 0 avatar

        @dal20402

        People won’t have to own autonomous cars. You’ll just call up a car when you need one. They’ll be circulating, and it will come quickly, take you where you’re going, and then go to someone else who needs transportation.

        There won’t need to be nearly as many cars on the road at once most of the time because people won’t have to be looking for parking spots. In a city of nothing but autonomous cars, the space now used for on-street parking could be taken up with an extra lane.

        (I’m not looking forward to autonomous cars, but just observing how I think things will be if they succeed.)

  • avatar
    stryker1

    It’s going to be really embarrassing when my iCar adapts to my driving/stopping heuristics, and when my mother in law borrows the car, it keeps trying to take her to Wendy’s…

  • avatar
    Scott_314

    I’ll take the negative view here. The Google car has $220K worth of equipment. A 10-fold price reduction would be required, but that likely isn’t possible. Autonomous cars need calibrated, sensitive equipment. They do not need more CPU power or hard drive space which do drop in price.

    And legal requirements may be insurmountable anyways. Who accepts the liability? I suppose an insurance company, but how much will they want for that? Even then, will they be willing to certify a car for all conditions? If only certified for some of the time, maybe on certain roads, etc., how will that be monitored and enforced?

    I’m saying it’s all feasible, but talk is cheap. Besides, hiring a taxi or taking the bus or train already provide the conveniences on the list.

    Let’s see what happens with Volvo’s demonstration in 2017 (100 cars only). But I don’t see this as inevitable.

    • 0 avatar
      dal20402

      Insurance companies will get on board real quick once they discover how much more accurate and less error-prone computers are than humans. Even if the computers screw up once in a while, and the insurance companies are held liable, their loss rates will be far lower than those incurred by human drivers.

      • 0 avatar
        Dan

        The liability problem doesn’t fall on the human drivers nor insurers. We’re off the hook when we’re not touching the wheel and so is the insurance we bought against our screwups.

        The automaker is on the hook instead. Their pockets run ten figures deep. Collisions with the autopilot on are an ambulance chaser’s wet dream.

        • 0 avatar
          dal20402

          Any risk that big will be insured. In the end, insurance companies are the specalists in managing risk, and they will handle big risks like this. If the drivers’ casualty insurers don’t pay, the automakers will insure the risk themselves.

        • 0 avatar

          > We’re off the hook when we’re not touching the wheel and so is the insurance we bought against our screwups.

          This is a nuance of the law not conducive to hasty judgement. Just because a piece of technology is involved doesn’t absolve responsibility.

          For example, radar cruise and auto parking exist now, and you can be sure to be on the hook if you run into someone with them.

    • 0 avatar
      JuniperBug

      The technology is only expensive because it’s new. Once the R&D cost is amortized and economies of scale ramped up, it’ll go the same way personal computers, webcams, fuel injection, cars themselves, and basically every other revolutionary technology went.

      At the rate technology advances and traffic increases, it would be silly to assume that autonomous cars won’t happen. It might not be tomorrow, or even 10 years from now, but they’re coming.

      In the long term, autonomous cars will make driving safer in the same way that giving control of commercial planes largely to computers has made them safer. The result will be fewer accidents and lower insurance costs than what we have now.

      • 0 avatar
        Scott_314

        I don’t know. How much cheaper have camera lenses gotten in the past two decades? Not at all.

        How much cheaper has a jet with autopilot gotten in the past two decades. Not much.

        Technology does get cheaper, and there are economies of scale. But look at fuel cell.

      • 0 avatar

        > The technology is only expensive because it’s new. Once the R&D cost is amortized and economies of scale ramped up, it’ll go the same way personal computers, webcams, fuel injection, cars themselves, and basically every other revolutionary technology went.

        Outside of Kurzweilian dumbasses “technological progress” isn’t really something which can be generalized.

        http://www.thetruthaboutcars.com/2014/02/qotd-toyota-not-tesla-as-a-force-of-disruption/#comment-2880161

        In this case the technology relies in part on computer hardware, which bodes well. It’s also fortunate that the production side relies in part on software (ie free to copy).

        Unfortunately because it works in SW, it also relies on algorithmic meta-reasoning which humans are somewhat limited in. We really don’t know what makes us “smart” and capable of accomplishing nuanced tasks, and mimicking the trappings of smartitude is often inadequate for the corner cases reality presents.

    • 0 avatar
      krhodes1

      @Scott_314

      Do you have a smartphone? How much would that have cost 15 years ago? Oh yeah, the technology to create an iPhone did not exist 15 years ago at any price. Do you REALLY think the same won’t happen with that $220K of sensors in the Google car? Seriously?

      This IS coming. Heck, a 2014 S-class is 80% of the way there NOW. Radar Cruise, Lane Keeping, auto braking, on and on. All that tech will be on a Corolla in 10 years. Just like stability control debuted on expensive cars, and ABS, and fuel injection, and on and on. We will not have fully autonomous cars all at once. It will happen in stages. Next will be essentially full autopilot on the highway. Eventually we will get to fully autonomous. It may take 25 years, but it is coming.

      I trust a computer to drive a car WAY more than a cell-phone clutching Soccer Mom. And so will the insurance companies.

      • 0 avatar

        > Do you have a smartphone? How much would that have cost 15 years ago? Oh yeah, the technology to create an iPhone did not exist 15 years ago at any price. Do you REALLY think the same won’t happen with that $220K of sensors in the Google car? Seriously?

        Hopefully you didn’t miss the common just above yours

        http://www.thetruthaboutcars.com/2014/03/ur-turn-autonomous-cars-are-already-here/#comment-2968529

        The problems with *entirely* self-driving cars isn’t the cost of the hardware, but rather because we *don’t know how* to make a smart enough computer yet. Don’t know how doesn’t follow Moore’s Law.

        It’s one of those goal where it’s a potential tossup whether or not sentient computers will come first.

        > This IS coming. Heck, a 2014 S-class is 80% of the way there NOW.

        This is a field where each % of progress becomes exponentially harder.

  • avatar
    Mandalorian

    I’m not really surprised. It was only a matter of time before lane assist, GPS, radar cruise control, blind spot detectors and all of the other tech came together.

    It will probably go something like this: Car drives on long straight highway stretches, human drives around town and does delicate maneuvering.

    • 0 avatar
      JuniperBug

      Given that there are affordable cars which can already park themselves, I think it’s not far out that they’ll eventually be able to do delicate maneuvers just fine.

    • 0 avatar
      dal20402

      Delicate maneuvers will be one of the first things to be computerized. My wife would be thrilled if a computer would back the car up the long, steep driveway to the garage, and that would be a super-easy task for a computer to handle.

      • 0 avatar
        redav

        I’ve heard Volvo is working on valet parking – get out of the car in front of the store and then send it on its way to find a parking spot. When you are done, you go to where you got out, give it a call, and it comes back to get you. You know that’s going to be a desired feature.

  • avatar
    CoreyDL

    I cannot imagine computer systems complex enough in cars to judge road conditions on every road accurately. If it’s snowy out, how will they know exactly how much ice is on the road, or when it was last scraped? How about if someone spun out a little and created a rut I might get stuck in? Is the system going to adjust for the fact that I have AWD, but did/didn’t put it in snow mode? How’s it know if I may have bald tires, and will take longer to get traction and make a left hand turn onto a busy street in bad weather? Is it going do drive around that snow drift which has partially gathered in my lane? Or will it stop me until it can make a safe lane change to the other (potentially) available lane, getting me stuck in the meantime?

    Say it’s windy and raining and it’s a garbage day. I’m going to work on a two-lane road, and there’s a garbage can in my lane, and a piece of a tree in the other, roughly parallel to one another. Does the computer stop the car, and flash a message that says CLEAR OBSTACLE? So now I must get out of my personal conveyance (where normally I would have probably driven around it somehow) in the weather and take care of it?

    Just seems too complicated, and unable to deal with all the various factors – let’s not forget needing to cover every road in America, even the private drives and other things which are not in GPS, or don’t have names, etc.

    • 0 avatar
      dal20402

      The computer system, once refined, is capable of judging road conditions very well, and — even more importantly — of reacting to new information far faster than you could.

      “How will they know exactly how much ice is on the road?” Sensors that can differentiate between reflections from ice, water, or pavement, reinforced by a combination of temperature, humidity, and weather information.

      “created a rut I might get stuck in?” The computer can see that rut and adjust for it.

      “snow mode?” “Snow mode” is just a different calibration for the AWD system. The computer can apply such calibrations on the fly depending on what it sees through its sensors and feels through the wheels.

      “How’s it know if I may have bald tires?” Because it will have already detected that the car slips more easily than expected given the weather conditions.

      “Just seems too complicated” Computers are able to handle “complicated” very, very, very well. It just takes time to develop systems capable of dealing with it.

      • 0 avatar

        Taking a look at the vehicles in the DARPA challenge are a good indication. Though sometimes they look more like DERPA machines…

      • 0 avatar
        CoreyDL

        Do you realize that any time someone says anything questioning the ability of a computer, you instantly reply, “It can do it.” You’re certainly sure of all the capabilities, and it sounds so easy! You must be at the top of the food chain at Googles Self-drive program.

        • 0 avatar
          dal20402

          I don’t need to know everything about the systems to know what I wrote above. I just need to know what has already been done and have a basic understanding of how computers work.

          • 0 avatar
            CoreyDL

            You cannot just -ass-u-me- it will all fall neatly into place that way.

          • 0 avatar
            dal20402

            No assumption necessary. Everything in my post has already been done. It’s really not a technological matter anymore — it’s a commercial and insurance matter, figuring out how to bring the technology to market. And the car and insurance companies will figure it out because the potential rewards are so great.

      • 0 avatar

        > “created a rut I might get stuck in?” The computer can see that rut and adjust for it.

        This isn’t necessarily true. There’s always going to be some set of conditions upon which the computer will decide to give up instead of doubling down.

        http://www.thetruthaboutcars.com/2014/02/analysts-peak-car-to-arrive-by-2020s/#comment-2871921

    • 0 avatar
      sportyaccordy

      Lol. If it’s snowy out, YOU don’t know how much ice is on the road, or when it was last scraped. YOU won’t know if someone created a rut you can get stuck in, until you’re stuck in it. My car can tell when conditions make for potential ice… it tells me in the dash. It wouldn’t be hard for a car to get live weather data and know what the conditions are before you do. And if conditions are bad enough that you can get stuck in a snow drift on the highway at speed, you probably shouldn’t be on the road in the first place. And I will bet an autonomous car linked up to a city road monitoring service would be able to tell you that as well.

      A well programmed computer can drive a car much safer than the average driver, period. More inputs, faster and more frequent reactions, the ability to utilize information a human can’t get (i.e. live weather reports, car to car communication to scan intersections and around blind turns, traffic conditions etc etc etc). It’s a no brainer. I think there are some legitimate gripes about self driving cars but the lack of capability of computer systems compared to human drivers is definitely not one of them.

      • 0 avatar
        CoreyDL

        ” And if conditions are bad enough that you can get stuck in a snow drift on the highway at speed, you probably shouldn’t be on the road in the first place.”

        Not what I said. I said two lane road. And who gives a frack about “probably shouldn’t” when they have to go to work?

        “And I will bet an autonomous car linked up to a city road monitoring service would be able to tell you that as well.”

        Right, because all towns have a road monitoring service. All of them. And they’re updated at 5:00AM in real time. Yep.

        As usual, you get up in arms and make a bunch of declaratory statements about whatever the topic at hand is.

        • 0 avatar
          sportyaccordy

          A road monitoring service could be implemented. There are already live traffic monitoring systems on most major roads.

          And you are the only one up in arms and making declaratory statements. There’s no hard proof that automated cars can’t work, or that computers can’t handle any of the things you are talking about. Plus you are assuming that all HUMANS can handle all of the things you are talking about, which I know not to be the case.

          • 0 avatar
            CoreyDL

            I’ve asked my laptop to drive me home. I’ll put it in the driver’s seat and see what happens. Will report back.

            “There’s no hard proof that automated cars can’t work”

            There’s no hard proof there aren’t aliens. Or that Noah’s Arc was real. What’s your point?

      • 0 avatar
        KixStart

        “Lol. If it’s snowy out, YOU don’t know how much ice is on the road, or when it was last scraped. YOU won’t know if someone created a rut you can get stuck in, until you’re stuck in it.”

        +1

        • 0 avatar

          > YOU won’t know if someone created a rut you can get stuck in, until you’re stuck in it.

          CoreyDL might not know much about computers, but his intuition that automated systems aren’t as resourceful as human is correct.

          If the thing’s stuck in a rut, what now? Ctrl-alt-del? Or write the essay with pen and paper instead?

          In the looooong run, the computers will get smarter, but this isn’t a trivial problem given even dumb people can be smarter than computers.

          • 0 avatar
            KixStart

            Even a small amount of snow during rush hour in the Twin Cities can result in hundreds of minor accidents, lane blockages and massive traffic delays.

            And that’s with the vehicles currently piloted by “resourceful” humans. I’m willing to give the idiot computers a try.

          • 0 avatar
            28-Cars-Later

            I’m surprised about Twin Cities drivers.

          • 0 avatar

            > And that’s with the vehicles currently piloted by “resourceful” humans. I’m willing to give the idiot computers a try.

            There are plenty of jobs which are seemingly simple, yet automating them entirely turns out harder than expected. Often the last few percents is far less trivial than its forebearers.

            We’ve been working on AI since forever, and thus far it can only accomplish rather trivial tasks with no breakthrough in sight.

            It’s far from a foregone conclusion that car will *entirely* drive themselves anytime _soon_.

          • 0 avatar
            dal20402

            If the car’s stuck in a rut, a well-programmed computer would try these steps (or something like them) in order:

            1) Try to drive out of the rut normally
            2) Use some stuck-in-a-rut techniques (rocking, etc.) that will be programmed into it
            3) Automatically summon help

            Basically, the same thing a human would do, except that the computer will be a bit more skillful.

          • 0 avatar

            > If the car’s stuck in a rut, a well-programmed computer would try these steps (or something like them) in order:

            Or even better, we can put a robot-human in the robot-car because it’ll figure out the context of the situation just as well.

            If the thing’s stuck because it just ran over someone, the best course of action isn’t going back and forth a few times more to get free.

    • 0 avatar
      JuniperBug

      Remember when people claimed that ABS systems couldn’t handle a variety of different conditions as well as a decent driver could? What happened there? Or to vehicle stability programs that can detect and correct slides far faster – and in more efficient ways – than a human ever could?

      If a human can detect a condition, a computer can be developed to detect it more quickly and reliably; it’s just a matter of developing the algorithms.

      • 0 avatar
        CoreyDL

        You’re just talking about components of the drive, not driving the entire car in all situations.

        • 0 avatar
          JuniperBug

          I’m also talking about things which were developed when computers were in their relative infancy, compared to now and the near-future, where they’re orders of magnitude more capable.

          ABS works with sensors which measure wheel speed and acceleration and can react to braking more quickly than humans. Vehicle stability systems use similar sensors and have algorithms which can also adjust the vehicle’s side-to-side trajectory relative to real-time road conditions better than a human can.

          Are you seeing the trend? Computers can measure where the car is going, where it’s supposed to go, and adjust its attitude better than humans can. You can poke holes all you want, but the writing is on the wall as to whether it’s humans or computers which can better control vehicles.

          How are 500 MPH, 100 ton jet aircraft steered again?

        • 0 avatar
          sportyaccordy

          The only thing stopping computers from doing all the driving is legislation. There are already cars on the road today with the capability to drive themselves. Not to mention the technology is in place to take over when situations overwhelm human capabilities. And of course there are the Google cars. So the idea that we are far away from the technology being feasible is flat out wrong. It’s already here and in heavy use every day.

    • 0 avatar
      krhodes1

      At first what will happen is the car will say “manual control needed” under these sorts of conditions. I don’t see this as a black and white situation. A 747 can fly itself from takeoff to touchdown, but we still pay a couple of pilots to babysit it. Self-driving cars will be no different. The technology WILL get better over time, but completely independent cars will be further out than sometimes drive themselves cars.

  • avatar
    doublechili

    The idea of only ever being a passenger is horrifying to me, but I understand the inevitability of this. The only question is the timing (“inevitable” could mean decades for full implementation). After MTs effectively disappear in all but rare models, the “automatic versus manual” designation will have a whole new meaning.

    Consumer preference will be the motivation for most, and insurance rates will be the means of convincing the recalcitrant. Eventually legislation will eliminate the wealthy recalcitrant holdovers.

    The refuges for enthusiasts? Virtual racing and indoor battery-powered race tracks for the masses, and racing country clubs for the well-to-do.

    Another step in the march towards safety-ease-comfort at the expense of individual freedoms. Inevitable and “better” but kinda sucks.

    • 0 avatar
      dal20402

      What freedom is being threatened here? If the government starts to legislate where we can drive our autonomous cars, or tells us we can’t drive our old cars on racetracks, then freedom will be threatened. But neither of those appears to be at risk of happening.

      • 0 avatar
        doublechili

        I’ve started to type a couple of different things, but honestly I’m at a loss as to how the question can even be seriously asked. How can not operating your own vehicle be considered as free as operating your own vehicle?

        BTW, “We can drive our autonomous cars” = great bumpersticker idea. Oh wait, in the future we won’t need bumpers….

        • 0 avatar

          There are many different possibilities, and they all stem from your interpretation of the word “free.”

          For instance, when I lived in Toronto, I didn’t use a car, instead taking a single speed bicycle everywhere (even in winter.) The money I saved was enough to purchase my first classic car. I sold that and bought my second. Freedom, for me, was financial: I took the pain and stench of cycling and converted it into the pain and stench of a classic car.

          For many, I suspect “free” will mean time. This can manifest itself in different ways, but—could an autonomous vehicle shave 10% off of your commute time? 20%? 50%? Given a more comfortable interior, could you do x more work tasks? Could you start “work” once you enter your car for the commute, your boss knowing full well you’ve got internet/phone access on your way into the office?

          Like others have said, you have to separate the idea of daily driving from the ideal drive—let the commute and chores fall to autonomy, and use the weekend for pleasure driving.

          • 0 avatar
            doublechili

            I interpret “free” as “free”. As in, I can operate my own motor vehicle today, and in the future at some point I may not be able to. Everything else is just semantic manipulation, respectfully.

            In other words, I can accomplish your versions of “free” voluntarily right now. If I am forced to adopt those versions of “free”, overall I am less free.

          • 0 avatar

            > Everything else is just semantic manipulation, respectfully.

            Dismissing his use scenario just because it doesn’t fit with how you choose to interpret a word is semantic manipulation by definition.

          • 0 avatar
            doublechili

            “Dismissing his use scenario just because it doesn’t fit with how you choose to interpret a word is semantic manipulation by definition.”

            Sometimes it’s not gray. Sometimes it is black and white.

            We were discussing a scenario where I would not be allowed to operate my own vehicle. I said that would be a loss of freedom. I would not be allowed, see? Suggesting that you might eventually find your enforced passengerdom “for your own good” really didn’t help with the whole loss of freedom aspect. In fact, I kind of bristle at that kind of thing. But maybe that’s just me.

            On a side note, my whole issue with the whole relative definition thing is that someday I might order a burger and get a chicken sandwich instead, and when I object they may insist that the chicken sandwich is in fact a burger.

          • 0 avatar
            dal20402

            Sorry, but driving your own automobile, as opposed to having a computer do it, is not a “freedom.” It’s a technique. The “freedom” which rises to the level of an actual human right is the freedom of mobility. And autonomous cars won’t threaten that in the least unless the government starts telling us where we can go with them.

            The eventual elimination of the human-driven car is in the same category as seatbelt and helmet laws — regulations designed to minimize the social cost of the freedom of mobility.

          • 0 avatar

            > Sometimes it’s not gray. Sometimes it is black and white.

            > might order a burger and get a chicken sandwich instead, and when I object they may insist that the chicken sandwich is in fact a burger.

            These are physical things, compare to “freedom” which is a nebulous cultural concept. Your “freedom” to live contradicts another’s “freedom” to shank you for some misdeed.

            I realize that in Merica it’s fashionable to consider it a noun which can be accumulated like money, but pretty smart people in linguistic philosophy have mulled over these issues, so unless you count yourself among them it’s best to believe they have the advantage here over the Billy-Bobs.

          • 0 avatar
            doublechili

            Dal, I’m not big on calling everything a “right”, so no argument there. It’s not about taking away a right, it’s about being less “free” on a sliding scale.

            You wrote: “The eventual elimination of the human-driven car is in the same category as seatbelt and helmet laws — regulations designed to minimize the social cost of the freedom of mobility.”

            I disagree with that. Seatbelts and helmets make us more safe as we autonomously operate our vehicles (BTW, I’d be fine with legislation telling us we have to wear helmets in passenger cars!). Similarly, traffic controls (lights, signs, speed limits, etc.) accomplish the same thing, along with facilitating the efficient flow of traffic (theoretically). But we’re still operating the vehicle. The discussion here is about going from autonomously operating our vehicles to being passengers on an amusement park ride, so to speak. We would no longer be “drivers”, and what we’re in would no longer be an automobile, it would be a pod.

          • 0 avatar
            doublechili

            Social, try saying “freedom is a nebulous cultural concept” to someone in jail (or in N. Korea). And, BTW, if the inmate happens to be, say, a compulsive eater, they may find that food limits would make them “free” of their compulsion, so I guess in the end it would be a good thing for them, like the road pods.

            If “free” is troublesome, let’s try “volition”. When my kids were a little younger they’d go on the kiddie car ride at amusement parks. They’d randomly turn the wheel. They’d honk the horn. But in the end the car just went along the rails at a set speed no matter what the kids did. And if any kids acted up the ride operator would stop the cars (seriously, think about that one).

            That’s what the future would be for adults. We’d punch in a destination, and that would be that. And articles and discussions on this website (if it still existed) would be about which cars looked coolest, which colors were best, which seats were most comfortable, which infotainment systems were best, etc.. It would basically be “Better Homes and Gardens”.

          • 0 avatar

            > if the inmate happens to be, say, a compulsive eater, they may find that food limits would make them “free” of their compulsion, so I guess in the end it would be a good thing for them, like the road pods.

            If you understand it’s a mere rhetorical game for certain words it’s unclear why not for the general category they’re from. To wit:

            > If “free” is troublesome, let’s try “volition”. When my kids were a little younger they’d go on the kiddie car ride at amusement parks. They’d randomly turn the wheel. They’d honk the horn. But in the end the car just went along the rails at a set speed no matter what the kids did. And if any kids acted up the ride operator would stop the cars (seriously, think about that one).

            Hopefully the kids went on the ride of their own volition. They must’ve considered it fun regardless of what the wheel did. Who are you to question that choice?

            If you wanted to play bumpercars, maybe the park has that available; if not, it seems only too Merican to sue them for denying “freedom of choice”.

          • 0 avatar
            doublechili

            Since “free” and “volition” didn’t work, how about “obfuscating”? Because this is really getting a little silly.

            By “volition” I did not mean relating to the act of getting on the kiddie ride, since that would be no different than the act of getting into one of the future transport pods. Rather, as an alternative to “free” since that was causing problems, volition was specifically referring to the exercise of one’s will upon the operation of the vehicle (hence, kids turning the wheel with no effect means no volition). And just to anticipate the next detour, punching in the destination isn’t the kind of “volition” I had in mind.

            Anyway, I may just have to surrender to the Linguistics 101-trained advocates of back door mass-transit. So, in sum: I can do a thing now. In the future I may not be allowed to do that thing. In fact, since the operation of my vehicle would be dependent upon systems beyond my control, my vehicle’s operation could be terminated (arrested) without my consent. Yet, I will be more free as a result. I’ll just repeat that 100s of times each day in between my visits to “The Truth About Pods” website.

          • 0 avatar

            > So, in sum: I can do a thing now. In the future I may not be allowed to do that thing.

            Well, time moves on. I wanna some slaves too but the gubmint hates my freedoms.

            In any case, sleep easy it prolly wouldn’t happen within your lifetime because old people can’t use computers and they vote. Personally I advocate online voting as the new poll tax but that’d never pass under the current oppressive suffrage regime–classic catch-22.

            > Linguistics 101

            Just btw, just because it’s called the linguistic turn doesn’t mean it’s purview of namesake dept.

        • 0 avatar
          sportyaccordy

          The prospect of autonomous cars does not automatically include the elimination of self-driving cars. Just from a practical point of view such a transition wouldn’t make sense. Nobody is going to take on the liability of retrofitting autonomous controls on normal cars, and nobody would stand for being forced to buy a brand new self driving car.

          Not to mention there is a small but vocal and dedicated minority of folks who enjoy driving, as well as several industries built up around them. Nobody wants a self driving 911 or Mustang or whatever. Used cars are much more robust today than they used to be as well… it will be decades before all the cars on the road today are off the road, and it will be decades before automated cars are available at a mainstream price.

          And of course, you are “frightened” with being a passenger, but I am “frightened” by inattentive, drunk, and beyond their use by date drivers… all of which could be eliminated with more flexible public transportation enabled by automated vehicles. Plus unless you drive something really old/analog odds are there is a computer handling some aspect of your driving at moments at which control/grip become life/death. So I’m not really sure what exactly you’re scared of; none of what you are saying jives with reality.

          • 0 avatar
            doublechili

            I actually never wrote that I was “frightened” or “scared” of being a passenger. I wrote that I was horrified of the idea of only ever being a passenger (ie., not driving anymore). I think that difference is pretty clear.

            I understand there would be a lengthy transition and that it could take decades for full implementation (I specifically stated that in my initial post). And, of course there would be a time where
            AUTO-mobiles and pods coexist. But do you really think the day won’t come eventually where automobiles are gone and it’s all pods?

            To put it another way, if X% of vehicles are pods and X% are automobiles, what do you think insurance costs would be on the autos? And once average folks are priced out of autos as a result, what do you think the media coverage would be like of accidents caused by wealthy auto drivers who can afford the insurance premiums? And wouldn’t legislation enacting a total ban on autos eventually come to pass, for the public welfare?

          • 0 avatar
            sportyaccordy

            Horrified and frightened are synonyms. Whatever. And yes I got that you meant not having the choice to drive yourself.

            I don’t think a day will come where there are no human driven cars. At least not in our lifetime. Insurance is a moot point. There is no reason for the cost of insuring human driven cars to go up… insurance costs are based on the price of risk; more automated cars on the road = a lower risk of a crash = lower insurance costs for everybody.

            Speculation about legislation is pointless. An equally compelling argument could be made that laws would be enacted to ensure some minimum amount of human cars on the road. After all, automated cars = less cars on the road… no sense owning a car if you could have 95% of the functionality with a fraction of the hassle & cost (if you don’t enjoy driving like most people). That means less people will buy cars and spells bad things for the auto industry, which has powerful lobbyists. So the convenient speculation WRT legislation is pretty pointless as well.

            There is nothing concrete to indicate the days of human driven cars is numbered… or at least numbered within our lifetimes.

          • 0 avatar
            doublechili

            I don’t know that “horrified” and “frightened” are true synonyms. I know people who are horrified at the concept of wearing white after Labor Day or eating iceberg lettuce, but they’re not frightened of it. :)

            You’re qualifying your points with “not within our lifetimes”. I don’t disagree since I really don’t know the timeline. Maybe it will be, maybe not. I do think the day of no autos/all pods will happen. And technology is moving so fast that it’s hard to say what might be possible in 30-40 years.

  • avatar
    JK43123

    I was totally shocked and horrified by the idea of a self-driving car. But I just recently returned from a 3 hour bus trip (charter so I knew everyone) and it was nice and relaxing letting someone else drive. I have always said I enjoy driving–and I do in new places–but do I really want to pay attention to the cathunk-thunk over the same bumps every day on my way to work? I say no, let the car drive while I relax before and after a long day.

    John

  • avatar
    blackbolt

    I can only imagine the thousands that would be made redundant if this comes to pass. Bus drivers, cabbies, truckers and such.

  • avatar
    Conslaw

    Bad things happen with autonomous cars. Proof: http://www.youtube.com/watch?v=4FTWzfpTvFs.

  • avatar
    kvndoom

    I don’t look forward to that day, if only for selfish reasons. Driving for me is like hunting or fishing for others.. it’s that one thing I do that I don’t want anyone messing with. It’s hard enough finding a decent car with a real transmission anymore.. good gawd the thought of having to shop for cars that still have a steering wheel!

    I see a big difficulty in autonomous cars and human-driven cars on the road at the same time. For the system to work perfectly, I see some network requiring feedback from all vehicles all the time.

    Could this lead to another “cash for clunkers” where cars with steering wheels are crushed to get them off the road?

    Then there are the impatient people. Obviously these new “future cars” will be programmed not to excessively speed (or not speed at all) in order to remain safe. Some people just hate going the speed limit. Some people lose it when they are late for work. Will they try to hack their cars?

    • 0 avatar
      sportyaccordy

      Again the existence of autonomous cars doesn’t mean the end of human driven cars.

      There are already autonomous cars dealing with all human driven cars now so if anything things will get easier and better integrated as autonomous cars become more normal.

      Govt wont be forcing anyone to do anything.

      Being that people can barely operate a computer, I find it highly unlikely that these same folks will be “hacking” their cars’ highly encrypted car control systems. If anything, speed limits for autonomous cars will be raised because they will be able to handle higher speeds more safely than a human would, and traffic would enable it.

  • avatar

    Worth mentioning there was a decent discussion on traffic effects in the last article related to robot cars.

    http://www.thetruthaboutcars.com/2014/02/analysts-peak-car-to-arrive-by-2020s/#comment-2867873

  • avatar
    05lgt

    Right after I learn the entirely predictable response cycle of the autonomous cars I’ll love goofing on them from my DRIVER’S SEAT.

  • avatar
    redav

    “For a driverless future to happen, two things need to happen. First, non-compliance with road laws and rising costs will make driving much more expensive—to say nothing of fuel prices. Second, technology will make it possible.”

    – There is a difference between autonomous cars (“driverless”) and no drivers in any car. Perhaps the above is necessary for the utter elimination of drivers, but for autonomous cars, the first requirement is unnecessary. There are thousands of drivers today that would gladly give up driving. In fact, on the radio this morning the DJs were talking about how they hated driving and if they were able to hire a driver, would never drive again.

    “The key to adopting driverless cars without outcry is to make drivers feel like they have a choice.”

    As noted in the example of the DJs, there will be no “outcry” regarding autonomous cars. They will be welcomed with open arms, regardless whether drivers feel they have a choice. The outcry will only come when people want to drive and are told they aren’t allowed, but also as noted above, that’s a different thing that “autonomous cars.”

    Given the title of the article, the fact is that autonomous cars are already here: self parking, lane departure control (that will steer for you), radar cruise control, automatic braking, turn-by-turn nav, etc. These systems do not yet work in all situations, but with each new product & each revision, they will do more until they do it all. It’s like the disappearance of cash. We will simply accept the increased conveniences until one day we realize we don’t need/use the old stuff anymore.

  • avatar
    GoesLikeStink

    Where will motorcycle fit in this new world? What will robocar do when a an old Shovelhead splits lanes next to it?

    • 0 avatar

      The computers could read motorcycles and start to lane drift, giving them more space. Hell, even activating the blinkers on the side the motorcycles should pass on…

      Since you guys haven’t yet protested the NSA, it may come to pass that license plates will be equipped with short-wave readable chips, which would be the perfect signature to detect those still driving / riding for themselves.

      • 0 avatar
        mcs

        What about equestrians? What about, oh, let’s say moose. Are you going to go around tagging wild life?

        • 0 avatar

          We haven’t turned them all into glue yet? Shit, let me write another story.

        • 0 avatar
          krhodes1

          How do you avoid moose now? If you are lucky you see them. Trust me, as one who lives in moose country you DON’T see most of them, so it is mostly luck. The computer has better eyes than you do. It will be able to see in the dark, and it will ALWAYS be paying attention. And just like now, some will hit moose anyway. But if the computer couldn’t avoid it, neither could you.

  • avatar
    mcs

    Okay, this is nice stuff to dream about and eventually we’ll get there, but here are some of the hard cold facts. Driving is dependent on a certain amount of intuition. The fact is that in many situations, by the time a hazardous situation is picked up by sensors, it may be too late to avoid. There’s the classic “ball rolling from between parked cars” scenario – a ball that rolls into the street may be followed by an unseen human, so even if it appears the ball will roll out of the path of the car, the human could suddenly end up in the path at a point where there isn’t enough space for the brakes to stop the vehicle. There are many other scenarios – like I live near a recreation area and I slow down when pedestrians, like a jogger or hiker, are present, but I’ll go even slower if a parent has a squirming toddler trying to break free from a parents hand. Anyone want to try to code that?

    So, one of the many problems is that we need better intuitive AI. We’re going to get there eventually, but it’s not nearly as easy as people think. I have co-developed a couple of successful collision avoidance systems for aviation, so I’ve been there. The intuitive part wasn’t as much of an issue, but outside of the controlled world of airports and the sky, it gets much tougher.

    Also, this sort of software requires zero failures. Zero. It’s not good enough to have one failure in a million miles. Not even 10 million vehicle miles. Zero. People die if you make a mistake.

    • 0 avatar
      05lgt

      It doesn’t need to meet “zero failures”. It just has to be enough better than what it’s replacing to justify the cost.

      • 0 avatar
        LALoser

        When leading a martial arts class, I always tell people that they do not need to be the best fighter around, just a little better than an attacker for a short period of time.

    • 0 avatar
      dal20402

      You don’t need zero at all. You need a rate decisively lower than the rate of human failure, so that the autonomous car is much cheaper to insure. And there is so much human failure in driving that it will be relatively easy to get there from where we are now.

      Also, a good chunk of the “intuition” you are talking about is learnable by AI with enough repetitions, and some more of it can be made up for by the massively faster reaction time of a computer.

      • 0 avatar

        > No assumption necessary. Everything in my post has already been done

        > Also, a good chunk of the “intuition” you are talking about is learnable by AI with enough repetitions, and some more of it can be made up for by the massively faster reaction time of a computer.

        Oh shiits, here come da computers experts to zoom and enhance their way to robot cars.

        • 0 avatar
          dal20402

          …says a talking robot with a shiny metal ass.

          • 0 avatar
            28-Cars-Later

            He’s also fun on the bun.

          • 0 avatar

            > …says a talking robot with a shiny metal ass.

            Fortunately I was programmed to bend pipe and own nubs.

            AI works largely via heuristics; the computer has no conceptual understanding of situations. Humans who work at that level wear helmets and post to TTAC from the short bus instead of drive any car. Intelligence may be an emergent property but what materializes from current ‘puter architectures can’t even out-recognize lower primates for the foreseeable future.

            So the questions is: can cars be programmed to drive safely in most all conditions with such a paucity of brainpower? If you know how to do that, quite your job and get a far better one at Google.

  • avatar
    brandloyalty

    First of all, driving is basically boring. The number of things people choose to do while driving proves that. If it weren’t socially unacceptable, we’d probably admit that we’d rather not drive most of the time. Second, operating cars will eventually be seen as an essential but dirty habit that should be minimized.

    Among the “many, many other things” that autonomous driving will bring, is something that will make the other changes look minor. Autonomous driving means there will no longer be any need for drivers of: taxis, transit vehicles, all delivery and transport trucks etc. There will be no need for taxis at all. There will be no need for trained drivers of ambulances and fire trucks. A whole class of jobs will be wiped out.

    One downside of autonomous driving is that some people will increase pollution by finding all sorts of reasons to move their unoccupied cars around.

    But this will unfold into another huge change. Private ownership and operation of cars will become relatively obsolete. Car cooperatives are already doing this. Most people don’t need their cars most of the time. It would save huge amounts of wealth, resources, parking and pollution to have commonly owned cars distributed and made available by predictive computerized systems. No more need to buy/store/insure or take the things to be serviced. Order the type you need when you need it. Probably you could even order up a performance car to take for a Sunday drive.

    Driver-controlled cars and autonomous cars will communicate with each other. The autonomous cars will know where the driven cars are and what they are doing. Networked cars will inform each other of threats such as ice. One car slips one wheel and the others in the area will know about it. Cooperatively owned cars means far fewer cars are needed and people can spend all that saved money on other things such as retiring early or working shorter hours. No need for garages or driveways. Any country adopting this technology will see a huge shift in wealth wasted on cars to things that improve quality of life.

    Soon cars will have “fly by wire” throttle, steering and brakes. Any car so equipped can be retrofitted with an autonomous driving system, and these systems will cost about as much as a fancy infotainment system. People who want to use bigger or fancier vehicles simply will be charged more.

    • 0 avatar
      sportyaccordy

      I think pollution would be less. I’d imagine many people would dump their personal cars and use public transportation or cabs, which would be made a lot more affordable with more sharing and automation.

  • avatar
    ajla

    The same automakers that today can’t get their ignition cylinders, radar cruise control, fuel tanks, and ECUs to safely function all the time are going to build us perfectly safe (and they will need to be virtually 100% safe) autonomous cars that can last 5-10 years and be priced to a level that most people can afford?

    It might happen, but I’ll be dead by the time it does. The best I’m ever going to see is a mega-fancy cruise control.

    • 0 avatar
      PrincipalDan

      http://en.wikipedia.org/wiki/Westworld

      Yes the idea of Autonomous Cars as being perfectly safe reminds me of “Westworld” – Where noting can go wrong.”

    • 0 avatar
      krhodes1

      They don’t have to be perfectly safe. They just need to be better than now. Which is an amazingly low standard. People SUCK at driving.

      • 0 avatar
        ajla

        “People SUCK at driving.”

        I believe that is something that will actually work against autonomous vehicles. Almost everyone believes they are a safe driver. It is the “other people” that need the robot cars.

        When a drunk or speeder kills someone there is an obviously preventable wrongdoing that caused the event. The thought is ” if I don’t drink or speed or go out at night around bars I will be safe”. If your autonomous vehicle has a system failure and you crash into a wall, what could you have done?

        Not to mention that the media isn’t going to go lightly on an crash in a robot car. An early fatality could put an end to the whole thing.

        That said, I don’t think we are going to go from a 2014 Versa Note to a fully automated transport pod. Things will be much more gradual.

  • avatar
    LALoser

    In PI I have a full time driver, and it is nice! The traffic is killer, so while he tends to that, I am sitting there drinking coffee and reading the paper. I get dropped right at the door. so parking and so on is not an issue. The added bonus is he goes back and takes the housekeeper to the store so that is another task made easy. He is off on Sundays, that is the only time I drive.
    Now that said, I know him, and I know he is professional and driving. It might be unsettling to hit gnarly traffic with just an echo in the car.
    Even in A/P transfer trains, you see the track and better understand what is occurring.
    They have to sell the security and then understanding will follow.

  • avatar
    danio3834

    Like completely pilot-less planes, completely pilot-less cars aren’t likely to ever come completely to fruition. While the technology exists and has existed for many years for fully autonomous cars to function under the most ideal situations, there still needs to be added redundancy of a driver.

    Unpredictable driving conditions can be accounted for in the software, but these systems are far from perfection, so they need to be able to default to human driver mode if there is a system fault.

    One of the biggest concerns that will need to be addressed as cars continue to become more autonomous is keeping the driver engaged so they have sufficient awareness to react correctly if needed. I’ve driven a 2014 Grand Cherokee for 4 hours straight, including through one of the most congested cities in North America during rush hour, using the adaptive cruise control system and didn’t touch the brake or accelerator once during that time. I did have my foot hovering *over* the brake pedal at certain points however. All it would have taken for the system to disable would have been snow and debris to cover the forward facing sensor, so the driver still needs to be ready.

  • avatar
    Kendahl

    The principal consequence of autonomous vehicles will be to turn automobiles into transportation appliances. There will be nothing left for the enthusiast. Who would spend the price of a Ferrari or put up with the discomfort of a Lotus when the software makes them drive like a minivan? And the whole point for an enthusiast is to drive the car himself, not be a passenger. To express it another way, when was the last time you spent most of a Saturday detailing your lawn mower?

    Autonomous vehicles should be able to speed up rush hour traffic. Traffic signals exist to resolve conflicts over right of way. When vehicles can resolve those conflicts themselves, there is no need for the signals or the delays they unavoidably cause. Instead of traffic waiting at a red light, think of vehicles going through the intersection alternately, slowing just enough to provide space for a crossing vehicle to slip through.

    Unless you have special needs or live in an area with a low population density, there would be no reason for you to own a vehicle. Just order one to pick you up at a designated time and location, tell it where to take you, and turn it loose after you arrive. The only difference if you did own it yourself would be to tell it to go park someplace until you needed it again.

    • 0 avatar

      > Who would spend the price of a Ferrari or put up with the discomfort of a Lotus when the software makes them drive like a minivan?

      Who would spend the price of a Rolex or put up with the discomfort of a heavy metal watch when piezo-quartz made it completely obsolete?

      It’s as if people buy Ferraris to drive them.

  • avatar

    Just a few notes on how computers works and relevance here:

    1. Computers, or rather the level we conscientiously understand reasoning, are very good at the general case, and very very terrible at exceptions. Google-cars can drive well in relatively straightforward circumstances (during sunny days around MV), and they know better than to embarrass themselves otherwise, on public streets at least.

    2. Some of the specific limitations such LiDAR implementations that don’t work in the rain can be overcome, but the reality is that it’s still a dangerous robot roaming on its own and just hoping for the best when it personally cares less about running 100 people over in case of malfunction. The risk tradeoffs here are not the same as a typical poor driver except maybe really old ones:

    http://www.cnn.com/2014/02/03/us/florida-elderly-driver-kills-three/

    3. The only sane solution is these first two problems is just shut the thing down whenever the sitch is sketch. That’s an entirely different *kind* of approach and issues than those facing a motivated sentient being capable of figuring things out. Computer are good at complementing human intelligence, and there’s a huge solution space unexplored by the industry before replacing drivers.

    To summarize, 80% solutions akin to advanced cruise control + steering + hud assist/warning are coming in the next 20 years, but nobody has the crystal ball about actual robot-cars in the near future.

  • avatar
    azmtbkr81

    Maybe I’m becoming a curmudgeon but with CAFE and platform sharing cars are becoming more homogenized and less interesting every year.

    Aside from a handful of unique models, nearly every new car is a regurgitated version of something else with the same unibody, suspension, engine, transmission, etc. In some ways I welcome the self driving car and hope it means a swift and merciful death to cars that are hardly worth driving anymore.

  • avatar
    PandaBear

    The biggest problem with driverless car is the detection of the road and its condition, in particular, the unexpected items that you cannot see.

    A child too small, a chihuahua, an broken piece of furniture, a water pipe fell out of the pickup truck in front of you that is still bouncing around, etc.

    Human eyes are extremely complex and sensitive, to a point that no professional camera can match. Our brains are extremely good at recognizing objects and patterns because we have 16 years growing it and hundred of thousands of years evolving it, and it will take a lot of training before a computer can be just as good in recognizing road conditions with just vision.

    Once you see something, picking the right decision (to stop, to slow down, or to ignore because it is a false alarm) is easy, and what a computer can do very well in.


Back to TopLeave a Reply

You must be logged in to post a comment.

Subscribe without commenting

Recent Comments

New Car Research

Get a Free Dealer Quote

Staff

  • Contributing Writers

  • Jack Baruth, United States
  • Brendan McAleer, Canada
  • Marcelo De Vasconcellos, Brazil
  • Vojta Dobes, Czech Republic
  • Matthias Gasnier, Australia
  • W. Christian 'Mental' Ward, Abu Dhabi
  • Mark Stevenson, Canada
  • Cameron Aubernon, United States
  • J Emerson, United States