By on August 15, 2012

Ben Klayman, Reuters’ Detroit-based crack car correspondent, wrote a very good feature on self-driving cars.  After interviewing many sources, he comes to the conclusion that “it’s been more than half a century since some of the first concept cars boasting self-driving features were presented to the world”  and that this probably will not change anytime soon. Even Eric Schmidt, CEO of Google and the staunchest supporter of the technology cautiously says that “self-driving cars should in our lifetime become the predominant way.”

1956 Firebird II – had autopilot

The answers Klayman received from experts range from  “My mental model of trust in technology is a Windows blue screen of death. That’s how much faith I have in PCs and computer systems,” said by Bryan Reimer, a research scientist at Massachusetts Institute of Technology’s AgeLab, to a despondent Bob Casey, curator of transportation at the Henry Ford Museum in Dearborn, Michigan, who seems to have accepted that the driverless car is coming and who is already bemoaning the past when real cars were still driven by real people:

“Part of the fundamental attraction of automobiles has been the actual driving of them. If you do away with that, then it really becomes an appliance … a toaster, a washing machine.”


1959 Cadillac Cyclone – had autopilot

BMW (“We will always be the ultimate driving machine”) does not buy the story, GM is all for it: “Once we have a car that will never crash, why don’t we let it drive?”

For me, the best part is Ben Klayman’s  factbox.  He actually went through the painful exercise of actually reading the study KPMG and CAR had prepared last week. He summarized the study’s findings, which read like the driverless cars will be the answer to society’s ills, stopping short of curing cancer and the common cold:

Possible consequences expected from driverless car

The good:

  • Automakers cut weight from cars and trucks as crashless cars do not need to be made with as much reinforced steel or as many safety devices like airbags. That would lower vehicle costs, speed up vehicle development time and boost fuel efficiency.
  • Automated cars would drive in tighter packs because computers would control their speed and spacing. That would mean smaller roads were necessary and result in the elimination of shoulders and guardrails, leading to a significant reduction in the $75 billion spent annually on roads, highways and other infrastructure.
  • With computers controlling the cars, driving would be more efficient and thus faster, leading to less congestion on the roads. Fuel consumption would decline and companies that rely on just-in-time delivery could reduce inventories even further.
  • Automated cars also would allow for the elimination of traffic and road lights in many cases. That would slash energy use drastically.
  • Driverless cars would mean a change in the way drivers are insured, and could even end the need for car insurance.
  • Crashless cars would mean auto repair shops see fewer damaged cars, meaning they would need to shift their business model to serving the aftermarket needs of existing cars that lack autonomous driving systems.
  • Steelmakers would have to adjust to a world where cars use less of their product.
  • Less expensive, driverless cars would open ownership to new audiences like younger generations or even the blind, but they also could lead to wider vehicle sharing that would slash global sales.
  • If vehicle sharing expanded, cars could be summoned as needed and people could pay for mobility services as needed instead of owning a vehicle.
  • Autonomous transportation could eliminate the need for and cost of high-speed trains.
  • Vehicle sharing could keep vehicles in more constant use, reducing the need for parking lots that take up a lot of land in cities.

The depends  on which side you are on:

  • Hospitals would lose more than two million crash victims sent annually to U.S. emergency rooms.
  • State and local governments would have to adjust to the loss of traffic fines, possibly reducing their police forces. Governments might seek to replace some of that lost revenue; perhaps with infrastructure usage fees.

The bad:

  • Lighter, easier-to-build cars could open the auto industry to new rivals using a model like Apple’s, where a company designs and markets a product but outsources its construction.
  • A connected, driverless car network would require security from hackers and would raise privacy concerns with many consumers.

P.S.: In a TTAC reader poll, 69 percent of the respondents thought driverless cars will revolutionize the industry,  31 percent thought they won’t.

Get the latest TTAC e-Newsletter!

32 Comments on “Self-Driving Cars: Don’t Hold Your Breath...”

  • avatar
    Da Coyote

    I’d love to have a driverless car, since the gummit has done a great job of making driving a pain in the Obama.

    However, when you think of a driverless car, I have one word which should scare you to death…..Microsoft. The “blue screen of death” – coming from the GM of software – would take on a completely new meaning.

    • 0 avatar

      777, 787s, A320, A340’s are fully computer controlled (with pilot override authority available on the Boeings) for a while now. The result is dramatically improved safety.

      In the period from 1962 to 1971 the fatality rate was 133 deaths per 100 million passengers. The rate from 1991 to 2001 was 20. From 2001 to 2011 the rate was 2.

      With that in mind, I’m not sure I share the same concern about the blue screen of death.

      • 0 avatar

        Yup, big aircraft are automated. But they still pay a pilot to sit in the pilot seat and – most importantly – monitor the plane and automation systems and be ready to assume control if anything goes wrong. And given the open skies, there’s usually several minutes and several alarms prior to any actual emergency manuevers the pilot may need to make.

        Conversely, your car won’t have a paid pilot, just you. And the response time to an upset won’t be minutes, but seconds. So you settle into your 45 minute commute, pull up your blackberry, ready your emails, and BAM you’re dead.

        Automating air travel is not like car travel.

      • 0 avatar

        “So you settle into your 45 minute commute, pull up your blackberry, ready your emails, and BAM you’re dead.”

        Do you think the rate would be more or less than 33,000 fatalities per year?

  • avatar

    The KPMG report reads like it was cribbed from Brad Templeton’s “Robocars” site:

    Which is recommended reading on the subject (he’s a true believer and is doing some consulting with Google on their project.)

    Because of the speed of development of cars (and the state of the technology) nobody should expect fully autonomous cars in their Chevy dealership by 2016. But they’re coming, and relatively soon: Google’s demo units are looking remarkably promising, and there is nothing inherently expensive about LIDAR systems that isn’t amenable to mass production (and huge price drops).

    If you push me to make a wild guess, I will say first fully autonomous car is in showrooms by 2020.

  • avatar

    I don’t see that “driverless” necessarily equals “crashless.”

    For one thing, the shift to self-driving cars will be gradual. There’s something like 160 milllion cars on the roads and less than 10% are replaced every year. When they are introduced, their rollout will probably be similar to that of the Prius or Volt, with just a couple per cent of the new car fleet being self-driving. Non-automated cars will be on the road for decades after the introduction of the first one. During this phase, you will need a car with good crash protection, whether it’s self-driving or not.

    And then, things go wrong. They’re going to go wrong at high speed and if we’re traveling in platoons, there’s going to be the probability that more than one car is involved.

    • 0 avatar

      Also more importantly, driverless cars at first will be specially adapted versions of human-piloted cars. It’s less expensive to leave airbags, other safety features in than remove them.

  • avatar

    •With computers controlling the cars, driving would be more efficient and thus faster, leading to less congestion on the roads. Fuel consumption would decline…

    •Vehicle sharing could keep vehicles in more constant use, reducing the need for parking lots that take up a lot of land in cities.

    One thing I wonder is, what is the net effect of fuel consumption given these two points? There are obvious fuel efficiency advantages with an autonomous car network. As car usage becomes more efficient and presumably more convenient, would car usage increase to the point of offsetting any efficiency advantages? With car sharing and the automated car-as-a-butler running around town for you, would cars be on the road MORE than they are today and dimnish the anticipated road capacity increases?

  • avatar

    Indeed, it may be technologically possible, but who would want to assume such liability?

  • avatar

    One thing I don’t see being addressed is the transition period where both types of cars exist on the road at the same time. As long as there are some “uncontrolled” drivers in the mix, the risk of accident still exists. Having an accident with cars that are travelling faster than most people would drive, in tightly knit packs, would result in mass casualties, aggravated by the plan for the new cars to have no airbags and be lightweight vehicles. You can’t strip down the robot cars and save on all that steel and safety equipment until there are no more legacy cars on the road, I would think. Unless you plan to build dedicated roads for the new vehicles…

  • avatar

    When Skynet comes alive, I’ll be on John Connor’s side. Driving around in a Marauder of some sort. Driving.

    No computer is gonna drive for me. Not now, not ever.

    And I’d love to see Google coming clean about the one single reason why they want us to stop paying attention to the road ahead: advertising.

    • 0 avatar

      Either you never fly, you’re one of the rare people that pilots his own aircraft everywhere or you have come to terms with not being in command of a vehicle in which you were riding.

      And about the time you’re 75 or 80, and your kids are trying to take away your car keys or maybe putting sugar in your gas tank, you’ll be looking upon this as a real benefit. I absolutely wish we had this technology available, now, because I know people who shouldn’t drive (young and old – but mostly old) and I know some people who have given up driving altogether or under all but the most favorable conditions because, although they need and want the mobility, they recognize the danger they may pose to others.

      • 0 avatar

        Ask me again when governments around the globe make self-driving cars mandatory. What do you think is gonna happen?

        It is freedom the thing I’m talking about. And by the time I’ll be 75 or 80, I will feel good about myself that I took the time to teach my kids to drive properly.

        For those who don’t want to drive, or who can’t drive, there’s always a taxicab nearby. Or public transportation. So why bother?

        Technology isn’t always the answer.

    • 0 avatar
      Robert Gordon

      So I take it your daily driver is a Ford Model T and you take great delight in twiddling the ignition advance and retard levers as you scoot along at a statly 20mph.

      If the above scenario is not true then you have already begun to outsource your driving.

      • 0 avatar

        Having electronics managing the engine is not the same as having a computer driving you around at speed. Outsourcing is a good thing when you do it so you can focus on what’s important.

        Ask any Forbes 500 CEO whether they’d oursource their core business. Driving is the core business here, not real-time dealing with ignition.

        And the electronics in my car fail all the time, but I don’t get myself killed for that.

      • 0 avatar
        Robert Gordon

        No, the core business is transportation. ‘Driving’ is merely an arbitrary term to group the subset of many discrete tasks that broadly relate to the operation of the machinery central to the aim of getting to the destination. Pressing a button that says ‘drive me home’ is just as much driving as the aforementioned juggling ignition levers. Whether you find it satisfying is another thing entirely, but suffice to say that sitting in a traffic jam or driving on a freeway is hardly satifying either.

        The car driving autonomously is just a logical incremental improvement to the motorcar that has given us electric starters over crank handles, fuel pumps over gravity feed, fuel injection over carburettors and so forth.

    • 0 avatar
      schmitt trigger

      “And I’d love to see Google coming clean about the one single reason why they want us to stop paying attention to the road ahead: advertising.”

      +1 You nailed it!!

  • avatar
    Mark MacInnis

    Yes, friends, Utopia is just a few years away.

  • avatar
    Mark MacInnis

    For some reason, that Cadillac reminds me of Madonna. Huh.

  • avatar

    I think the first driverless cars won’t be cars, they will be long-distance trucks. Not only will trucking companies save the cost of a driver, they will get around on-duty requirements and the technology can be introduced in the easiest possible environment: Interstates.

  • avatar

    The real appeal of the driverless car, when you peal it back, is that it is basically a form public transportation for people too snitty for public transportation. The convenience of a bus/subway/tram without having to mix with the plebs. I’m sure Harry “the tourists stink in summer” Ried, then, would be all for it.

    And with that, I’ve got no use for it.

    • 0 avatar

      I’m having a hard time understanding how the driverless car relates to Harry Reid but maybe it’s just your primal need to b1tch-slap a Democrat whenever possible.

      I’ve used public transportation many times, all around the globe, but I generally don’t use it around my home and neither using it nor eschewing it is because of my Marxist political leanings, it’s because of convenience, cost and scheduling. I can walk to work faster (45 minutes) than I can get there by bus (90 minutes). I have real sympathy for the people that actually have to rely on mass transit around here.

      As it is, mass transit doesn’t go everywhere, all the time, it makes stops I’m not interested in, which lengthens the trip, and it isn’t dispatched the moment I decide I want transportation. Taxis can do that but they are prohibitively expensive for regular use. An autonomous car solves a lot of these problems. If I own one, it’s entirely at my disposal and if we start using shared pools of them, I won’t have all the fixed costs of a car but I’ll have lots of the benefits by using pool vehicles. I imagine I’ll be able to reserve them in advance for a fee, save money if I can do flexible scheduling and maybe participate in dynamic vehicle sharing for a reduced rate. I won’t have to pay to park a pool car.

      In all likelihood, most families will use some mix of private and pooled cars to maximize convenience while minimizing cost.

  • avatar

    All of the self-driving vehicle tests I have seen are in parts of the US with relatively mild climates (mostly CA, NV, and AZ).

    How are these systems going to handle the rough and tumble weather the rest of the country gets like the hard winters we get in the northeast? What will happen when the camera/LIDAR lens gets blocked with snow or obscured by rain? Are they going to put defrosters/wipers on them? Will it even know to slow down during inclement weather, or will it keep trying to plow along at the speed limit during a blizzard? How will it maintain lanes when the lane lines (or even the road itself) is no longer visible because of snow drift?

    Also, I wouldn’t count on a reduction in repair work. Will the system swerve to avoid a pothole or road debris (even it it means temporarily leaving the lane), or will it just plow through it? I can see a lot more bent rims and alignments coming out of this.

  • avatar
    Chicago Dude

    As a computer scientist with some amount of training and experience in AI techniques, I am excited to see self-driving cars on the roads (when they are ready). They will seriously change the urban world as we know it when used as taxis and paratransit vehicles. As soon as you no longer have to pay the driver (who is usually a horrible driver anyway) fares will drop, accidents will be reduced, traffic will be reduced, etc.

    Now, as for private ownership and operations… The benefits listed above hinge on one single question. Will the self-driving car be able to self-change a flat tire? If not, are we going to expect a blind man or a child to change it? If not, how will we ever be able to do away with road shoulders and other safety-related road features? Is the car going to sit in the driving lane, blocking traffic until someone can come to the aid of the car? Of course not.

    Self-driving cars will be great. But let’s not get carried away with it.

  • avatar

    There’s no such thing as a crashless car, thus the entire write up of pro’s and con’s is false. Automated driving will reduce and potentially eliminate driver error, but that’s not the same thing as “crashless”. If anything, driverless cars can exponentially increase the severity of a crash by failing to respond in the correct manner (due to a failing in original programming). The total number of crashes may reduce, but the severity may increase. And there will still be crashes.

    I work in manufacturing industry and we deal with automation failure all the time. ALL THE TIME. ALL. THE. TIME. So don’t count on a crashless anything.

  • avatar
    Robert Schwartz

    I agree that self-driving cars are inevitable. I don’t think that this decade is a reasonable time frame. I think that we would be looking at a generation for such cars to reach the point of ubiquity and perfection that would be needed before we ban manually controlled cars fro using the streets and highways.

    That said we are seeing the first developments down the path toward self-driving cars. Not, just Google’s experiments, but elements like adaptive cruise control and automatic parking are becoming widely available. Soon they will be universal and their capabilities will expand.

  • avatar

    I have no doubt driverless cars are the future as the technology matures and proves itself in all aspects of driving due to the numerous advantages machines have over humans (although as others have said, the assumption that there will be no accidents and hence no need for modern safety equipment and road shoulders is either naive or short-sighted).

    However, I also have no doubt that even after the technology has been proven, it’ll take at least one or more generations for people to accept them due to the large economic and societal impacts. In addition to the industries mentioned in the article, I don’t know why anyone would ride a bus if they could easily borrow a friend’s car (driverless cars would basically be like having a personal driver) or presumably hire a taxi for a fraction of the current cost.

    In a way it will be similar to the problems we’re having now in adopting digital media. The technology is already in place to stream or download virtually any piece of audio, video, or written content ever made but (at least for now) access to the content is hopelessly tied up over digital rights/intellectual property rights.

  • avatar

    They will pry the steering wheel from my cold dead hands.

  • avatar

    As much as I love cars and driving I’ll be the first one to admit it’s not the ideal transportation system in the same way someone who loves horses wouldn’t try to convince you that saddling up is the best way to get to work. A pool of shared autonomous vehicles driving in sync with each other would be a much better use of our limited resources, be it time, money or environmental resources.

  • avatar

    Driverless cars can’t happen fast enough in my opinion. By my observations 80% of the driving public does their best to detach themselves from the driving process as it is. Better to get them fully out of the driving loop so I don’t have to worry about them merging into me while they text on their cellphones.

  • avatar

    First, one big thing about comparing FAA software to MS-bloat is that the FAA can come into the software house and demand an explanation for every single line of code a programmer ever committed. Vastly different from what is in windows, office, or iDrive.

    I can’t belive the list above. If automated cars come, it will be driven by:

    Old people who know they can’t drive, but won’t give up their freedom. Nearly all are in the latter, the former takes awhile.

    Soccer moms aren’t quite the political force they are said to be, but they all vote and would love to have an electronic chauffeur available.

    Chronic drunk drivers aren’t a political force at all (and MADD will fight tooth and nail to prevent them from going safely to and from bars), but will certainly cut down the 30k/year deaths.

    Drafting. Somehow they the article never mentioned the gas mileage benefits to drafting. Notice the olympic peloton at ~35mph? Any idea how much more important it would be at >70? Also allowing such close following would pretty much mean a break from endless and expensive building of new highways. People won’t care about this until after it is done, then won’t know how they lived without it.

Read all comments

Back to TopLeave a Reply

You must be logged in to post a comment.

Recent Comments

  • SaulTigh: I’m super glad to see Murilee making that effort. The mileage is one of the best parts of his...
  • SaulTigh: It feels like the crypto people are doing that today. I sometimes feel like a fool for...
  • mcs: @Dahlquist: No, it’s more than just sniffles. Maybe you should read about it.
  • EBFlex: “Omicron is more contagious. you should read about it.” Yeah and consists of the sniffles and a...
  • dal20402: Those are the classic Mercedes gauge faces, essentially unchanged (except for a subtle change from a...

New Car Research

Get a Free Dealer Quote

Who We Are

  • Adam Tonge
  • Bozi Tatarevic
  • Corey Lewis
  • Jo Borras
  • Mark Baruth
  • Ronnie Schreiber