Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!

 



Forgot your password?
typodupeerror
×
AI Transportation

Are We Ready for Driverless Trucks? (cbsnews.com) 313

Two million truckers move 70% of America's goods. But hundreds of thousands of their jobs could be disrupted away, reports Jon Wertheim on the CBS news show 60 Minutes, in "a high-stakes, high-speed race pitting the usual suspects — Google and Tesla and other global tech firms — against small start-ups smelling opportunity."

One of those startups is TuSimple, and their company's chief product officer points out that an AI driving system never gets distracted or falls asleep at the wheel: Chuck Price has unshakable confidence in the reliability of the technology; as do some of the biggest names in shipping: UPS, Amazon and the U.S. Postal Service ship freight with TuSimple trucks. All in, each unit costs more than a quarter million dollars. Not a great expense, considering it's designed to eliminate the annual salary of a driver; currently around $45,000. Another savings: the driverless truck can get coast-to-coast in two days, not four, stopping only to refuel — though a human still has to do that...

Jon Wertheim: How far are we from being able to pick up the specific cars that are passing us? "Oh, that's Joe from New Jersey with six points on his license.

Chuck Price: We can read license plates. So if there was an accessible database for something like that, we could...

Test Driver Maureen Fitzgerald: This truck is scanning mirrors, looking 1,000 meters out. It's processing all the things that my brain could never do and it can react 15 times faster than I could.

Most of her two million fellow truckers are less enthusiastic. Automated trucking threatens to jack-knife an entire $800 billion industry. Trucking is among the most common jobs for American's without a college education.... Sam Loesche represents 600,000 truckers for the teamsters. He's concerned that federal, state and local governments have only limited access to the driverless technology.

Sam Loesche: A lot of this information, understandably, is proprietary. Tech companies wanna keep, you know, their algorithms and their safety data — secret until they can kinda get it right. The problem is that, in the meantime, they're testing this technology on public roads. They're testing it next to you as you drive down the road...

This discussion has been archived. No new comments can be posted.

Are We Ready for Driverless Trucks?

Comments Filter:
  • Intermediate steps (Score:5, Interesting)

    by Nidi62 ( 1525137 ) on Monday August 24, 2020 @06:03AM (#60434943)

    Before we have autonomous trucks I think we will have semi autonomous trailer convoys. Basically 3, 4, 5 trailers attached to motorized, driverless cabs linked to a lead, manned cab leading the whole train. Can even have the semiautonomous cabs collecting data for you for the eventual switch to fully autonomous. But even then you will have 1 driver doing the work of 3 or 4.

    • Only about 23 states permit "triple tow" (two trailers) and that would have to change in order to enable such road trains. It would make a lot more sense to run more freight rail, though. Trucks suck.

      • The extra trainers are not connected to one truck. Each trailer would have a truck, with the front one being manned and the rest being self driving in convoy mode.
        • by Nidi62 ( 1525137 )

          The extra trainers are not connected to one truck. Each trailer would have a truck, with the front one being manned and the rest being self driving in convoy mode.

          I would imagine it wouldn't even be a "truck", since it wouldn't need a cab. Just a motor, wheels, communication/data/monitoring equipment, and a body to hold it all in. And as a bonus, once the convoy gets to the facility the autonomous trailers can split off from the convoy and use GPS to independently move to docks for loading/unloading, since those pathways would be known and static (take pathway A to door 1, pathway B to door 2, etc).

        • I see. They don't need a human for that, though. Also, it's not safe. Following distance isn't just about human reaction time. It's also about issues like parts falling off, blowouts, other vehicles getting in the way, etc. We have a superior technology for having trailers closely follow trailers already. It's called rail. You can load shipping containers or entire trailers onto rail cars, we do both of these things already. We should do more of that instead of playing stupid games with trucks.

    • Not going to happen on most roads. But I can see it being a possibility for those Australian trucks that are 3 or 4 trailers hooked up to a monster truck that travel for days across the outback.

    • by Osgeld ( 1900440 )

      at that point why not use trains, cause its got the same problems, my stuff is in the middle car and its at a depot

    • by eth1 ( 94901 )

      Before we have autonomous trucks I think we will have semi autonomous trailer convoys. Basically 3, 4, 5 trailers attached to motorized, driverless cabs linked to a lead, manned cab leading the whole train. Can even have the semiautonomous cabs collecting data for you for the eventual switch to fully autonomous. But even then you will have 1 driver doing the work of 3 or 4.

      This would only work on highways. Around here, you have some traffic lights that will turn yellow again before a single heavy semi can get all the way through an intersection, let alone the 4 more following it. So now you have several trucks running a red light to keep up with their convoy, or the front of the convoy has to stop in the street to wait until the others catch up. Either way will create havoc.

      You could only do this if you can magically materialize the convoy on a highway, and disperse it before

  • Seriously, we to change our taxation to deal with more automation, as well as to encourage bringing all the automation into the nation.

    Hopefully, Tesla gets smart and pushes for allowing automated trucks to run on federal highways from 2100-0530 on clear nights and with no construction/obstruction on the route. Doing runs between warehouses would be perfect start.
    • by burtosis ( 1124179 ) on Monday August 24, 2020 @07:59AM (#60435225)
      Then we need to start a universal basic income. Everyone uses the excuse of “The buggy whip makers just got other jobs!”, But there will come a day when we will have strong AI or capable enough weak AI and cheap androids such that there will be no jobs for 99% of people. Why pay meat to work for far more money than you can purchase a machine to work 24-7-365 (366 on leap years)? This may be 20 years off, 50, or even 100, but if a UBI system isn’t in place by then it’s going to be too late.
      • The reason, as we see today, is that "meat" is cheaper. Those machines require a large up-front cost and costly maintenance.

        We have fruit picking machinery, yet we pay labourers (ie cheap immigrants) to pick fruit.

        We have automated warehouse picking machines, yet we pay labourers (ie cheap immigrants) to pick crap off warehouse shelves.

        We have automated car wash machines, yet we pay labourers (ie cheap immigrants) to wash cars by hand.

        I'm just beginning to see a common factor here!

    • In 50 years you won't be ALLOWED to drive on public roads. It will be self driving only, with central routing and override of vehicles to optimize traffic.
  • by Generic User Account ( 6782004 ) on Monday August 24, 2020 @06:13AM (#60434963)
    The apostrophe is not an accident, is it?
  • They can set up the digital highways and do it. Its not expensive, you just cameras, local positioning, 5g and remote control override capabilities, plus large shoulders, etc. But the "final miles" would be trillions of dollars for changes, the first and last 4 miles of routes not to modern recievers.

  • Are driverless trucks ready?
  • Why worry? (Score:5, Insightful)

    by AndyKron ( 937105 ) on Monday August 24, 2020 @06:36AM (#60435029)
    Companies can't get enough drivers as it is so I don't think current drivers need to worry. Besides, automation took all the jobs away back in 1978. I read it in a magazine once.
    • Re: (Score:3, Insightful)

      by Powercntrl ( 458442 )

      Oh, I'm sure we'll invent more telephone sanitizer jobs for the unemployed truck drivers. If Covid-19 has taught us anything, it's that a good portion of the public are employed in "non-essential" jobs which only exist to enable them to participate in the economy.

      • If Covid-19 has taught us anything, it's that a good portion of the public are employed in "non-essential" jobs which only exist to enable them to participate in the economy.

        Most non-essential service jobs are there to provide services that people want but don't want to do themselves. Those are legitimate jobs. Jobs like long haul trucking exist because we spend our subsidy money on roads instead of rail. Those jobs exist to enable other people to participate in the economy, as a means of producing pork. Because we choose to spend the money on roads in order to benefit big oil, big auto, big rubber and so on, we have to have those trucks, so we have to have those truck drivers

  • I trust driverless trucks with driving on highways, but not on streets. Streets are where we have more issues like snow not being shovelled (remember, LIDAR doesn't work with snow), tighter spacing in roads, humans, etc.
    The thornier issue I forsee will be insurance, as normally you can blame any crashes on the human. Even if they weren't driving, they should have been watching out for mistakes. Eventually once the states/countries and insurance weigh in, you'll see driverless insurance.
    Where I see things go

    • Where I see things going once we figure out the insurance issue is driverless to a hub either on the highway or with a direct exit on a highway then human on the streets.

      The obvious thing to do is to run from truck stop to truck stop, which is a reasonable place to match up drivers with trucks. But it really makes much more sense to run freight on rail, and match trucks up with cargo at rail depots. We should be outlawing rails-to-trails projects, and instead refurbishing rail lines for freight carriage. Instead, government introduced a thing called "Rail Banking" where rails can be removed without losing right-of-way in some situations, so that in theory rail service can b

  • by argStyopa ( 232550 ) on Monday August 24, 2020 @07:09AM (#60435075) Journal

    First, you have to have the technology capable of actually doing this. I think it's irrefutable that we have the basic ai driving tech in place for driving down a clear dry level road in Mountain View CA. But of course, there's a strong Pareto principle involved. Nearly ANYONE can drive down a clear dry level road in Mountain View CA. It's the weather, traffic, obstructed views, unpredictable others, pedestrians, animals, equipment failure, etc that make driving...interesting.
    I'm not entirely convinced that we're there yet for that last 20% of driving. (TBH, that's the point of failure for humans as well; and there are a lot of questions about how we determine what is an acceptable level of 'good enough' for humans, as well - for example for seniors whose abilities are fading.)
    All of those are magnified in complexity and difficulty when your vehicle is 80,000 lbs and 60' long (36t and 20m for our imperial-impaired friends).

    Secondly, this being America, we need a LEGAL regime conceptually prepared for these vehicles. This is a complicated enough question for humans to resolve, but if a AI car hits a pedestrian, who's liable? The carmaker? The passenger? The software developer? Worse, of course, is if the AI has to make a trolly-problem decision; effectively the result will have to be coded by SOMEONE who is then directly responsible for a death, even if that meant fewer other deaths and was the "best" choice.
    We are totally nowhere near that yet.

    While I understand that there are compelling amounts of money behind this in a pressure that doesn't yet exist for cars, maybe we want to start this tech out on TRAINS?
    They have comparable (if a few orders-of-magnitude bigger) issues with weight and length but substantially simpler operating environments - no route decision, no oncoming traffic, nearly no pedestrians, and repetitive routing that would be ideal for 'training' the AI in a very clear set of "weight A, weather B, location C = speed X" rules.

    Once that's all working, let's implement that in SIMPLE real world car situations before going to the hardest route-driving challenge of trucking.

    • if a AI car hits a pedestrian, who's liable? The carmaker? The passenger? The software developer?

      If Ubers example of BSing the data by darkening footage before releasing it to the public, trying to hide the fact they had most or all of the safety systems disabled because it stopped the vehicle too much, and blaming the emergency take over driver - staring for hundreds of hours doing absolutely nothing then supposedly a human is able to take over in a split second, has taught me anything it’s that NO ONE is going to pay.

    • by drinkypoo ( 153816 ) <drink@hyperlogos.org> on Monday August 24, 2020 @08:19AM (#60435297) Homepage Journal

      While I understand that there are compelling amounts of money behind this in a pressure that doesn't yet exist for cars, maybe we want to start this tech out on TRAINS?

      We already have self-driving trains. The technology is completely irrelevant to self-driving trucks because everything about the operating environment differs. It is however yet another good argument for using more rail and less road. Taking the steering out of self-driving makes the whole thing dramatically easier.

      • But those trains require a human operator anyway. At least freight trains do. If it's really self driving, it shouldn't need a human. Otherwise it's more of an emergency braking system than a self driving system.

  • Great, one of the last relatively stable, well paying, and high demand blue collar jobs now goes poof.

  • Now that the trucks drive themselves, the Teamsters will have to go back to handling teams of horses.
  • Well how secret can they keep after an deadly crash in court?

  • Yes, we are ready, but I don't know if the technology is ready. I live in an area where two long-haul interstates merge briefly where they cross. We have wrecks here CONSTANTLY. At least a few a week, sometimes a few per day, and they almost always involve tractor trailers.

    One thing many people don't realize is the truck driver turnover rate is insane. Do you know what the annual truck driver turnover rate was in 2018? 98% That is almost unbelievable, but it also explains why there are so many wrecks - most

    • Well, that's super misleading. That's only at large fleets. Aren't most truck drivers working independently or in small fleets?

      When I worked in shipping, I saw the same 15 drivers every day for a couple of years.

  • by Osgeld ( 1900440 )

    what kind of dumbfuck question is that? no we are not ready to let barely functional hardware barrel down the interstate with damn near 100,000 LBS of dead weight behind it

  • by nicolaiplum ( 169077 ) on Monday August 24, 2020 @08:33AM (#60435387)

    Self driving trucks are going to happen whether the Teamsters, or anyone else, wants it or not. It is inevitable that when trucks can drive themselves they will do so, and it is inevitable that long distance routes - on predictable, grade-separated roads - will be automated first. That means long-distance truckers will be rare - first they will be driving the lead truck in a platoon with automated followers, later the trucks will be entirely autonomous.

    The Teamsters can take this gracefully or they can fight it and lose, and their members will suffer more. The example they should follow is that of containerisation of docks and shipping in the 1960s. Most unions and dock workers realised that the mechanisation of docks and the replacement of break-bulk stevedores with container crane operations was inevitable - and they worked with dock owners and operators to preserve the jobs of the existing workforce as far as reasonable, and to ensure that the existing union organisations were not displaced when the existing job types disappeared.

    The Teamsters can either accept this, ensuring that workers who will no longer be driving trucks are given alternative careers or alternative roles n the transportation industry, or they can fight a rear-guard holding action until there are large scale job losses and the entire Teamster union becomes irrelevant and is suppressed.

  • I could see an uptake in truck hijackings. divert the truck , unload it and sell the goods, maybe hack the computer so it thinks it is still on the road.
    Lots of interesting problems to solve. What do these things do if it is cloudy and GPS is not accurate. Do they 'keep trucking'. How and where do they recharge/ refuel? The downside is whole industries being destroyed, not just drivers jobs, but truck stops etc. Upside, is food prices should go down because the drivers never need to sleep and deliverie

What hath Bob wrought?

Working...