SoylentNews is people
SoylentNews
Waymo officially expands self-driving effort into trucks
Waymo, Alphabet's self-driving company born out of Google X, is seen by many as the leader in the field of self-driving.
After focusing on autonomous passenger cars to soon launch a self-driving ride-hailing service, the company is now expanding the effort to trucks. The company has been known to have been working on a truck program since last summer, but they confirmed it today in a blog post.
[...] Now the program is expanding to Atlanta, Georgia, which they will make the home of Google's logistical operations. From there, Waymo will ship cargo to Google's data centers. They say that you will be able to see Waymo's blue trucks on the road as soon as next week as part of the pilot program
Also at TechCrunch, Ars Technica, and Reuters.
(Score: 2, Funny) by cocaine overdose on Sunday March 11 2018, @04:42PM (6 children)
(Score: 0) by Anonymous Coward on Sunday March 11 2018, @05:16PM (5 children)
And no one will use their qualcomm anymore. We can go back to calling in on every stop and hoping the dispatcher picks up at 3AM. Filling out our paper logs and all will be well.
If you think the drivers will have a say in what happens you have not paid attention. They will just put 'loaders' on the trucks and send them along with the truck. No driver needed. The drivers will not even be involved. Plan on it. Your job is going to be automated away. Probably within 15 years.
How do I know this? I converted your paper logs into digital ones. I am the one who made sure your hours were straight so your terminal did not get shut down. I interacted with your bosses. They *only* keep you around because you are necessary. The second you are not you are gone. Trust me they are planning on getting rid of you as soon as they can. A computer can drive 24/7/365. A driver can drive 11 per day. That is 13 where the truck is not rolling. If it is not rolling it is not making money. They can do 1.5 to 2x with auto drive vs a driver. There is 0% chance a company is going to turn that down.
(Score: 2) by Runaway1956 on Sunday March 11 2018, @05:30PM (4 children)
A "loader", you say? WTF for? Automated forklifts will eliminate any need for a "loader".
(Score: 1) by Ethanol-fueled on Sunday March 11 2018, @06:51PM (2 children)
So when those forklifts inevitably tip over because their physics and control feedback engines were written by Indian coders, will they have a robotic voice that will squawk, "I've fallen...and I can't get up?!"
(Score: 0) by Anonymous Coward on Monday March 12 2018, @12:50AM
nah, the coding will be by Boston Robotics. When they start to trip and fall, they will just do a flip and land on their wheels again.
(Score: 3, Informative) by schad on Tuesday March 13 2018, @06:22PM
No. It turns out to be more cost-effective to leave them where they are and route the other automatic forklifts around them. Once all the loading docks are hopelessly blocked by broken auto-lifts, you simply bring in the robot bulldozers, level the entire facility, and rebuild it from premade modular components shipped with self-driving trucks.
(Score: 0) by Anonymous Coward on Sunday March 11 2018, @08:29PM
At first it will be loaders. Then nobody. As you correctly point out it will be fully automated. I do not see autoload of trucks, for now. As there is a bit of an art to it, and the problem is well into the NP hard domain. But yes eventually. Automatic driving is more along the lines of path following and do not hit things in the way (I am simplifying). More in the NP complete set of problems. We are using hyper maps to basically brute force the problem. So even auto driving is not a 'solved' problem.
If the loads are uniform then autoloading is a fairly simple problem and just a matter of getting the pallets the right size and distribution. In a LTL situation or something like a department store (walmart) the loads are not uniform and take a bit of work to keep the trucks from tipping over. Yet still pack them dense enough as to not waste space, weight and fuel.
My point was they will put loaders on the trucks at first. For something such as a day cab driver. That is basically what they do. They load/unload/drive. But if they do not have to drive the price the pay is significantly less. Think along the lines of your local KFC getting a shipment of chicken from its local distributor. There needs to be someone there to sign off on the bill of lading. Surprisingly enough people do steal from the backs of trucks all the time (including chicken). There will become a class of 'drivers'/'loaders' who live in their trucks. They do not stop ever. They just live out of the truck as if it were a home. The rules would allow it. As the driver is no longer in charge of the truck and can go off duty/sleeper. In the back of the truck. Pair drivers used to do this all the time with 5x4 or 4x4 hours (cant anymore because of the new rules). But if the guy is not driving he can basically be 'off'/'sb' all the time. The new rules discourage that behavior because of driving. But take driving out of the mix and as soon as they can work around it, they will. Like I said I made the rule system for trucks for a long time. I know the tricks. They are all part of my test cases. I probably even know a few most drivers dont.
Like you point out if they can crack loading automated then yeah the loaders are gone too. I do not see that quite as soon as auto driving as they are very different problem spaces. One is basically the traveling salesman problem the other is probably the bin/box packing problem (in at least 5 dimensions). https://en.wikipedia.org/wiki/Packing_problems [wikipedia.org] People have been working on that problem for a long time. As uniform packing is easier and faster for load/unload. But it is not in any way, solved. Also like I pointed out one of the problem spaces is not even a computer science thing. It is a regulatory thing (many of them). Which require people to be involved. That is changeable but not always due to political forces.