Follow Slashdot stories on Twitter

 


Forgot your password?
Close
typodupeerror
×
Transportation

Self-Driving Trucks Begin Mail Delivery Test For US Postal Service (reuters.com) 103

Posted by msmash from the marching-forward dept.
The U.S. Postal Service today started a two-week test transporting mail across three Southwestern states using self-driving trucks, a step forward in the effort to commercialize autonomous vehicle technology for hauling freight. From a report: San Diego-based startup TuSimple said its self-driving trucks will begin hauling mail between USPS facilities in Phoenix and Dallas to see how the nascent technology might improve delivery times and costs. A safety driver will sit behind the wheel to intervene if necessary and an engineer will ride in the passenger seat. If successful, it would mark an achievement for the autonomous driving industry and a possible solution to the driver shortage and regulatory constraints faced by freight haulers across the country. The pilot program involves five round trips, each totaling more than 2,100 miles (3,380 km) or around 45 hours of driving. It is unclear whether self-driving mail delivery will continue after the two-week pilot.
This discussion has been archived. No new comments can be posted.

Self-Driving Trucks Begin Mail Delivery Test For US Postal Service More

Self-Driving Trucks Begin Mail Delivery Test For US Postal Service

Comments Filter:

  • Taxis and mail already?

    • Re:I thought it was not ready? (Score:5, Interesting)

      by jellomizer ( 103300 ) on Tuesday May 21, 2019 @02:48PM (#58631950)

      I think this particular run between Phoenix and Dallas is a rather easy drive. Something I actually think would be a better job for a Train.

      Heck for a lot of our long travel over 30 miles, we can probably be better off having a train with a flatbed, we park our cars on it, and ride the train to our city hub, which we drive off and go the last few miles.

      • I think this particular run between Phoenix and Dallas is a rather easy drive.

        Agreed; let's see some true tests in the Northeast, in the snow.

      • we can probably be better off having a train with a flatbed, we park our cars on it, and ride the train to our city hub

        The really interesting thing is that some time ago in the UK, a study was commissioned to determine the best way to increase the efficiency of the rail network. They studied the various routes, and took into account both long-distance and shorter travels.

        The conclusion they came to was that the best way to improve the system - something like four times better than the next available option - would be to pave over all the rail tracks and turn them into roads.

        Yeah, I know, citation needed. Meh, I don't care i

      • Heck for a lot of our long travel over 30 miles, we can probably be better off having a train with a flatbed, we park our cars on it, and ride the train to our city hub, which we drive off and go the last few miles.

        That depends on your tolerance for loading/unloading times for vehicles. You wouldn't use a flatbed carriage. A standard length carriage would only hold four vehicles. You'd use a double deck extended carriage that can fit twelve. Loading/unloading is done by driving through the train from one end with little bridges between carriages in order to traverse the gap. If the don't decouple the autocarriages then you'd have to back all the way out in order to disembark, which people are dumb so the cars would ha

    • I thought it was not ready?

      Hence not one but two people required. Rest assured their required 'interventions' will not be publicized.

    • the pace is much, much faster than we've been lead to believe. Heck, mail should be easier. Same routes every day. If there are problems with the road quality you can fix the roads enough to make it work. Whatever it costs to fix the road gets saved in wages.

      The only Trouble is mass layoffs for what are in many places the only middle class jobs.

      • Re: (Score:3)

        by lgw ( 121541 )

        The only Trouble is mass layoffs for what are in many places the only middle class jobs.

        Driver is a working-class job. Heck, airline pilot making $300k is a working class job. The difference between working class and middle class is not what the job pays, but the nature of the work.

        Perhaps you meant "the only well-paying jobs"? But the guys who make good money driving trucks are the guys who own the trucks. That won't change with automation.

        • Driver is a working-class job. Heck, airline pilot making $300k is a working class job. The difference between working class and middle class is not what the job pays, but the nature of the work.

          Negative, the difference in class is your lifestyle. In referring to economics however, class is mostly about income, not what you do. If you make 20k as a manager, then you won't be able to afford a middle class location and lifestyle. If you drive a truck and make +100k then you certainly will if you don't blow your money.

          • Re: (Score:2)

            by lgw ( 121541 )

            Negative, the difference in class is your lifestyle

            These are social classes. There are working-class jobs, and there are middle-class jobs, and that's not directly about pay. There generally aren't upper-class jobs. The bell curves of pay for working-class jobs and middle-class jobs overlap quite a bit.

            In referring to economics however, class is mostly about income, not what you do.

            In economics, it's better to talk about income quintiles, or similar mathematically-defined boundaries. What your talking about is political propgana, not economics (OK, OK, there's not really much difference between these two these days, so maybe I conced

    • Re: (Score:2, Interesting)

      by Anonymous Coward

      Depends on your definition of 'ready'.

      Most people talking about it being not ready are talking about being generally ready to tackle all of the same conditions that human drivers deal with. That is to saying, driving with icing on the sensors, poorly marked roads, driving in locations that are GPS denied or having issues with GPS signals (mountain driving or dense urban driving has issues with this), driving in areas that have no or inaccurate mapping data, and conditions in which the rules of the road are

      • driving with icing on the sensors,

        If we want to be comparable to human would this be driving with ice on the eyes or ice on the windows?

  • Going postal has a whole new meaning now (Score:1)

    by Anonymous Coward

    lol

  • But to have a safety driver and an engineer seems excessive. A normal truck has just a driver, sometimes a passenger for when they are training the new guy. Couldn't they just Certify the Engineer to handle the safety driving on the truck?

    • for just a five round trip trial?
    • He isn't union. :) Actually odds are the driver is a postal employee and the other is from the truck company.
    • which is kind of a pain. When you've got a CDL there's all sorts of extra requirements on your regular DL (at least there are in my neck of the woods).

    • Re:I understand this is a trial run... (Score:5, Insightful)

      by Zmobie ( 2478450 ) on Tuesday May 21, 2019 @03:10PM (#58632096)

      Not to state the obvious (at least I thought it was obvious), but I'm sure the engineer is there to monitor specific performance metrics and make tweaks to the system in case of issues. It's extremely common during test stage of any system for an engineer to be present... While doing this, the engineer is likely not going to be paying attention to driving, more paying attention to how the system is trying to drive which is a key distinction.

    • On the serious side, the driver is there to watch the road, the engineer is there to watch the telemetry.

      On a non-serious side, one person is there to watch Netflix, the other person is there to stop the first from watching Netflix... (or was it Hulu?)

    • But to have a safety driver and an engineer seems excessive

      That's basically a vote of "no-confidence" in the technology.

    • A normal truck has just a driver, sometimes a passenger for when they are training the new guy.

      Depends on the destination.

      DOT regulations require that a single driver not drive more than 11 hours at a give stretch and can be "on duty" for no more than 14 hours a day. Now a truck sitting around for 10 hours isn't making any money. So most trucking companies will have two drivers if the destination is more than what can be driven in 11 hours.

  • From what I've heard from folks from that industry, the driving jobs are actually the more sought-after ones, from a quality of life perspective.

    The office/sorting jobs are the boring/grueling ones.

    Makes sense to automate what you can I suppose.

    It's still going to take a lot of time to automate it, since the percent of deliveries that just fit the letter/postcard format, or mail boxes that don't conform to standards will mean that almost all routes need some human attention for a good while.

    But given the na

    • Re: (Score:2)

      by flippy ( 62353 )

      From what I've heard from folks from that industry, the driving jobs are actually the more sought-after ones, from a quality of life perspective.

      The office/sorting jobs are the boring/grueling ones.

      Makes sense to automate what you can I suppose.

      It's still going to take a lot of time to automate it, since the percent of deliveries that just fit the letter/postcard format, or mail boxes that don't conform to standards will mean that almost all routes need some human attention for a good while.

      But given the nature of the rules forced upon the USPS for de-facto pre-funding retirement for any new employees - I can imagine wanting to spend a lot to preclude that in any way you could.

      Ryan Fenton

      They're not talking about mail delivery here, just the hauling between distribution centers. Once the big truck is loaded, driving it from the Phoenix USPS center to the Dallas one or vice versa doesn't involve the actual delivery to the addressee, which I agree will involve human attention for a good long time to come.

    • Truck driving isn't for everyone, that is why there is a lot of demand for the job, and their pay for what is considered a blue collar job, is still decent.
      Driving hundreds of miles a day defiantly can get boring and grueling, however you need to make sure you are alert as to not get in an accident. Depending if you work for yourself or for a company, your amount of work freedom varies. If you are working for a shipping company, you can't say who your customers are, and where to go. You go to where your co

    • It's still going to take a lot of time to automate it, since the percent of deliveries that just fit the letter/postcard format, or mail boxes that don't conform to standards will mean that almost all routes need some human attention for a good while.

      Over a decade ago, our post office tried enforcing a rule stating that post boxes had to be a minimum height above the street (due to the purchase of new delivery vehicles). While we did raise ours, even just along our street I notice the boxes are at varying heights. Then there is the multitude of different styles - plus a few security boxes.

      I realize this story isn't about end delivery - but I would be surprised if they solved that in the next decade. And frankly, I get very little useful "snail mail" any

      • The Post Office has been pushing neighborhoods towards communal boxes, which would ease the task of making them autonomy-friendly by a lot, since they would design and install them.

        Not that the test in this article is about that.

    • Re: (Score:3)

      by Zmobie ( 2478450 )

      Lot of sorting is actually already automated at USPS facilities (at least the large ones). Can't speak for the individual branches as I always dealt in industrial scale sortation systems, but typically in a good system 95%+ should be automatically handled by a large system.

  • Better ways to transport bulk cargo than truck (Score:5, Interesting)

    by Solandri ( 704621 ) on Tuesday May 21, 2019 @02:49PM (#58631958)
    If you *really* want to transport bulk cargo without a driver, rather than figure out how to make trucks drive autonomously, why don't you just put it on a train? The rails guide the train even if the conductor falls asleep, so it'll be a helluva lot easier to make trains autonomous than trucks. The only question then is if the additional cost of transferring the cargo between truck and train at the source/destination city is less than the cost of making trucks autonomous. Chances are it is - the container shipping industry has done wonders for reducing onload/offload costs.

    Economically, trucks are 3x more expensive than trains [cbo.gov] (table A-4, page 35). The U.S. just has an obsession with truck transport because the Interstate Highway System subsidizes truck transport. About 50% of the cost to build and maintain highways comes from fuel taxes on cars, while about 90% of the wear and tear on the highways is due to trucks. If you corrected that, trains would suddenly become the preferred method of long-distance land transport again, as they are in the rest of the world.
    • The US already transports loads of bulk cargo via train.

    • Re: (Score:3)

      by Zmobie ( 2478450 )

      Yes and no. In a hub and spoke model train routes are incredibly efficient to use for regional transportation. However, in a point-to-point system trains lose a LOT of efficiency simply because it is so expensive to build out/use/schedule rails to suit their needs because they are so fluid. You can get a better idea of this with airline company comparisons. The best example I use is American vs Southwest because they are headquartered near each other, face similar economics, etc., but AA uses a hub and

    • Intermodal freight is older than most Slashdotters.

    • Re: (Score:3)

      by Kjella ( 173770 )

      If you corrected that, trains would suddenly become the preferred method of long-distance land transport again, as they are in the rest of the world.

      Meh, trains are cheap when you ignore the absolutely ridiculous investment cost to build a separate infrastructure exclusively for trains with very low tolerances for curves/hills and the fact that you're fucked whenever there's a closure for maintenance or it's blocked or there's signal errors or there's a train set failure. Not even high speed rail and freight trains like to play on the same tracks, unless you want to get very little of each. If you could have autonomous electric trucks shuffling goods ar

      • Re: (Score:1)

        by Anonymous Coward

        "Somewhat more efficient?"

        Even the most advanced semis in tip-top repair are only half as efficient in terms of gallons per ton of cargo per mile. And most semis on the road are not even close to operating at peak efficiency.

        Increased Intermodal (i.e. containers transferred form tucks to rail and vice versa) is the best path to go using existing infrastructure. Shifting even 10 percent of existing all-semi routes to intermodal would be estimated to save 1 Billion tons of fuel annually without any substantia

        • Re: (Score:2)

          by Kjella ( 173770 )

          Increased Intermodal (i.e. containers transferred form tucks to rail and vice versa) is the best path to go using existing infrastructure. Shifting even 10 percent of existing all-semi routes to intermodal would be estimated to save 1 Billion tons of fuel annually without any substantial capital investiture and would reduce substantialy the annual wear on the interstate highway system.

          While that might be true the actual driver is like 0.01% of the cost of a freight train and self-driving wouldn't really add any new flexibility or capability, so if they prefer semis over trains today it's not going to change that. Self driving trucks on the other hand would massively change road transport to make it much more attractive. Of course it doesn't preclude a hybrid model but I think the threshold before you start building new rail would become even higher, when it's already high today. Here in

          • The cost of the driver isn't what's pushing for automation. It's the lack of drivers and a desire to run a truck 24/7 so you can operate a fleet of fewer vehicles and lower your capital investment costs.

        • And water transport (think canal/river) is significantly more efficient than train. So let's dig canals everywhere. Ignore the fact that they where everywhere and got ran out of business by the railroads. I agree, though, that intermodal is better way to go than long-range trucks. And the short-range truckers get to see their families more often.
          • minimalism is significantly more efficient than water transport. 'less is more' - no you don't really need that new pair of shoes or a 4k video of cat jumping.

      • A) Your argument that the build out would be too expensive is rather moot, given that the US already has the largest rail network in the world by far [wikipedia.org]. It’s nearly twice the size of the next closest country, China, and nearly as large as China’s, Russia’s, and India’s (the next three countries’) rail systems put together. Moreover, the reason most Slashdotters seem to be unaware it’s so huge is because the US rail network is almost exclusively used for cargo, meaning the

    • About 50% of the cost to build and maintain highways comes from fuel taxes on cars, while about 90% of the wear and tear on the highways is due to trucks.

      This is not true. The figure for wear and tear from trucks is closer to 99%. Road damage increases with the fourth power of the weight of the vehicle, so roads that are built strongly enough to handle 50-ton semis barely even notice the passage of passenger cars. To a first approximation, trucks should pay all of the highway taxes.

      If you corrected that, trains would suddenly become the preferred method of long-distance land transport again, as they are in the rest of the world.

      Indeed. Correctly allocating highway taxes would make trains the carriers of choice.

      • Re: (Score:2)

        by lgw ( 121541 )

        To a first approximation, trucks should pay all of the highway taxes.

        hen a truck delivers groceries to the store, everyone is using the roads. Highway taxes are silly, since effectively all road wear is from logistics that benefit everyone.

  • An unexpected cost of automation is going to be wage inflation. If automated things like self-driving cars require less maintenance then wages for maintenance workers will continue to rise to offset the reduced labor. This will be interesting to see played out in different industries. Is it really cheaper in the long run to automate things? How many industries will be unable to pivot away from automation if the price becomes too much for the market to bear?
    • that's not how it works at all.

      Wages don't go up when there are fewer jobs. Instead competition for the remaining jobs increases and wages go down. That's because supply of labor is outstripping supply of jobs. What will be "interesting" is what happens when 1/4 of the workforce is suddenly (on a global timescale) unemployed and unemployable. And by "interesting" I mean as in "Interesting Times". The last time we had high unemployment like that we got a World War out of it.
      br. Just out of curiosity wh

      • that's not how it works at all. Just out of curiosity where did you get this notion that having fewer jobs and more job applicants results in higher wages? Did it come from a news source, pundit or website?

        It's a hypothesis I have. This is based on how wages are for other industries where maintenance is done heavily. Look at jobs like auto repair, HVAC, plumbers, etc. As the technology becomes more complex and specialized, the jobs require more training, and higher skill workers demand higher wages. Since maintenance jobs are essential and maintenance becomes less routine with improved quality of the systems, there is less labor to be done but it's more highly specialized, resulting in higher wages, I believe

        • Do you really believe _plumbing_ is more complex and specialized today? Seriously?

          Shit flows downhill.
          Payday is friday.
          Keep your fingers out of your mouth.

          Plastic pipe makes most things much easier...Hope it works out better than last try.

          Gotta find another explanation for the plumber shortage. It's not that it's difficult. In CA the construction plumber trade is all Mexican (who work cheap), but maintenance plumbers are hard to hire. Could be that the plumbing company rates are high, but pay is

        • Re: (Score:2)

          by bws111 ( 1216812 )

          In 1948, a carpenter made about $2.09/hour. Today it is $23.86. In 1948, an auto mechanic made about $1.60/hr. Today it is about $21.50.

          I think you are confusing the 'labor rate' that is charged to the customer with the actual salary paid to the mechanic, plumber, etc. The 'labor rate' is a misnomer, it is actually the 'cost of business rate'. That is, you add up all your expenses (all employees (including benefits, vacations, etc), rent, equipment, insurance, utilities, advertising, etc, and some prof

    • Re: (Score:2)

      by Zmobie ( 2478450 )

      The economics still apply with wage elasticity. The companies will only pay to a certain point where they can maintain profitability. Wages will likely consolidate for the remaining jobs to some degree, however newer jobs will be created as a result of the automation that also are factored into the equation. Not to mention there are measurable benefits outside of reduced labor costs to consider, such as fewer mistakes, more strict application of business rules, and lower employee benefit cost (in that se

  • Are 'safety driver' and 'engineer' required to meet fed limits on daily hours etc?

  • The USPS is not just another company. It is part of the United States Government. And it has unions that are approved by Congress. Good luck replacing the drivers. Just a hint of the difficulties of replacing postal employees with self-driving trucks: https://www.apwu.org/news/dept... [apwu.org]. Also remember that there are tens of thousands of veterans in the Postal Service.

    • The employees will continue to be paid the same, they'll just get to sleep during the ride. Don't think they'll have any objection to that.

  • the carrier will deliver mail to the houses then when a few blocks away from the truck, whistle and the truck comes to them instead of them having to walk back to the truck.

    • Re: (Score:2)

      by Rolgar ( 556636 )

      Couldn't you setup a phone app with location detection, sych with the truck, and the truck just follow the mailman when he's out of the truck? In my neighborhood, that would save the mailman walking from the door to the curb at every house, or parking and returning to the truck after he's walk several houses up the block. When there's a package, the truck is close by without the carrier having to keep the try near by the extra effort. Over time you might end up being able to drop a percentage of the carrier

  • Haven't even gotten self-driving cars ready to roll yet.

    Self-driving cars are one thing, but the notion of a Self-Driving large truck is terrifying.

    Large trucks are dangerous enough when they're not being controlled by a computer.

    So.... What will happen when there's no human driver, and something goes majorly wrong?

  • I would have thought that they could make an arrangement to have a separate slow lane on the highway- maybe with a 40mph speed limit between 11pm and 3am, low enough to make drivers want to keep out of it, then post cameras and sensors along the roadway. These things would help develop an autonomous trucking system along trial routes. The "safety driver" could be back at base watching a video feed, so long as there are a few supervisors and automated failsafes.

  • It is finally here!

    "A safety driver will sit behind the wheel"

    Oh, um. Carry on.

  • This will probably worry a few (million) people
  • So I would imagine that one thing that keeps young people out of the trucking industry is the belief that in a few years it will be taken over by robots and they will need to train for another job. I certainly know I wouldn't bet on trucking as a career path. So by trying to alleviate the problem, autonomous trucks may inadvertently be contributing to it.

    • That's the whole idea behind the FUD - to create a shortage that necessitates our $1M autonomous truck.

Slashdot Top Deals

Genius is ten percent inspiration and fifty percent capital gains.

Close