Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Self-driving trucks begin mail delivery test for U.S. Postal Service

reuters.com

Post reply on HN
Print view View original

81–90 of 169 posts

Re: Self-driving trucks begin mail delivery test for U.S. Postal Service

#81

Post by esilver »

SmellyGeekBoy wrote: post #75
esilver wrote: post #66

Earlier quoted context omitted.

It would be wholly dependent on the load and the technical implementation. In my view the monitor would be a driver with standard knowledge and experience who could drive if the autonomous system encountered an edge case or for last-mile, i.e., from the freeway to the shipping or receiving facility. Under those circumstances a load traveling ten hours with six hours of loading and unloading time would require two hou…

It's not like the monitor will be sitting there playing Angry Birds all day - they will be required to be 100% focused on the road so they can intervene at a moment's notice. Personally I'd find that more stressful than just driving myself.

The arrangement I described above assumes no need for active monitoring for long-distance driving. It is the most monotonous and least technically-complex part of the process and the driver would be free to do other things during it.

It’s not the same process as the autonomous urban driving being attempted by Waymo et al. Driving from Phoenix to Dallas on the highway is far simpler than driving around Phoenix.

Re: Self-driving trucks begin mail delivery test for U.S. Postal Service

#82

Post by 0xffff2 »

linuxftw wrote: post #67
0xffff2 wrote: post #52

Earlier quoted context omitted.

>Semi's can't react as quickly to changing conditions as a Tesla. That's true whether there's a human driving or a computer. The only relevant distinction for a semi or a Tesla is whether the computer can react faster and make better decisions than a human driver would.

Wrong. Driving a commercial vehicle requires a license that is more difficult to maintain, and for good reason: It's much harder to safely operate than a standard passenger vehicle. Tesla and other 'Autonomous Driving solutions' are missing parity in the 'contextual awareness' space IMO. Safely operating a commercial vehicle is very much about 'contextual awareness'. Car cuts you off unexpectedly? Easy, slam the brak…

>Wrong.

How petulant.

> Car cuts you off unexpectedly? Easy, slam the brakes in a Tesla. Truck? Crash and kill the people in front of you.

Again, how is this different between a computer and a human driver? As far as I can tell, the hardest part of self-driving is accurately seeing the world. Once you see a car, it's fairly easy to make the algorithm pessimistically assume that any car that can cut it off is likely to do so. If the car manages to cut you off anyway, you're going to "crash and kill people in front of you" either way, except that you're probably going to kill people over so slightly less often with a computer that makes the decision to brake 100ms earlier.

Re: Self-driving trucks begin mail delivery test for U.S. Postal Service

#83

Post by ghaff »

dexen wrote: post #79
ghaff wrote: post #56

Earlier quoted context omitted.

There's a chain of hotels (Yotel) that is known in part for its robot that will stow luggage the hotel is holding for you. In practice, the staff mostly stows your luggage like every place else because the robot doesn't have anything like the throughput you need at busy times.

The flipside to it is, during slow hours you still have the robot at the ready to help the odd guest. The robot just sitting there generates virtually no costs (bar the minuscule amount of electricity to maintain situational awareness), while at the same time freeing humans to do more important tasks. Which may very well be family and social life.

Maybe? There are always people on duty anyway and the robot contraption probably takes up way more expensive urban space than any cost savings it brings. I'd be pretty sure the robot is far more about having a hip urban vibe than any sort of ROI.

Re: Self-driving trucks begin mail delivery test for U.S. Postal Service

#84

Post by lm28469 »

esilver wrote: post #51
YeGoblynQueenne wrote: post #40

>> A safety driver will sit behind the wheel to intervene if necessary and an engineer will ride in the passenger seat. This is how you know a vehicle is self-driving: there are two people sat in the front.

The in-vehicle monitor will likely be a standard part of autonomous and semi-autonomous trucking for the foreseeable future. A monitor can be paid less than a driver, as they’re doing less work, and can respond to onsite issues more promptly and accurately than could a remote one.

> A monitor can be paid less than a driver, as they’re doing less work

People working full time still need to live you know, even if it's "less work". Last time I checked the "amount" of work doesn't really correlate with the pay too.

You can't tell a professional truck driver "Ok man, you'll be in the same seat, on the same road, for the same amount of time but you get $500 instead of $1500 because it's less work, oh and also if there is a crash and you weren't 100% focused you'll be liable for the damage"

Re: Self-driving trucks begin mail delivery test for U.S. Postal Service

#85

Post by freehunter »

esilver wrote: post #78
ddoolin wrote: post #73

Earlier quoted context omitted.

> as a bonus he gets to unload the truck. Literally unload? Does this mean something else? I wouldn't see this as a bonus.

Some drivers have to unload their trucks depending on the load and the receiver. For multiple LTL (less-than-truckload) loads the driver might stop at multiple facilities and unload a handful of pallets at each. This would be more likely for smaller-format discount retail like, e.g., the Dollar Store or a restaurant supply chain.

I have family that works for a food distributor where their drivers get to restaurants, unload the truck, stock the items in the fridge/shelves, and take additional orders from the customer before they move on to the next delivery.

Re: Self-driving trucks begin mail delivery test for U.S. Postal Service

#86

Post by Symbiote »

dexen wrote: post #79
ghaff wrote: post #56

Earlier quoted context omitted.

There's a chain of hotels (Yotel) that is known in part for its robot that will stow luggage the hotel is holding for you. In practice, the staff mostly stows your luggage like every place else because the robot doesn't have anything like the throughput you need at busy times.

The flipside to it is, during slow hours you still have the robot at the ready to help the odd guest. The robot just sitting there generates virtually no costs (bar the minuscule amount of electricity to maintain situational awareness), while at the same time freeing humans to do more important tasks. Which may very well be family and social life.

You could also just have lockers.

Re: Self-driving trucks begin mail delivery test for U.S. Postal Service

#87

Post by esilver »

jsymolon wrote: post #70
esilver wrote: post #51

Earlier quoted context omitted.

The in-vehicle monitor will likely be a standard part of autonomous and semi-autonomous trucking for the foreseeable future. A monitor can be paid less than a driver, as they’re doing less work, and can respond to onsite issues more promptly and accurately than could a remote one.

> ... can be paid less than a driver, as they’re doing less work ... Major West: And the monkey flips the switch. -- Lost In Space, 1998 Instructions can be, push the big red button if anything happens.

If it were as simple as pushing a button a monkey could fit the bill. Or maybe a pigeon. [0]

The monitor would be in place to manage the last-mile driving and any other technically-challenging aberrations on the route. Human skill and judgement would still be required, just for much shorter periods.

[0] https://en.wikipedia.org/wiki/Project_Pigeon

Re: Self-driving trucks begin mail delivery test for U.S. Postal Service

#88

Post by esilver »

freehunter wrote: post #85
esilver wrote: post #78

Earlier quoted context omitted.

Some drivers have to unload their trucks depending on the load and the receiver. For multiple LTL (less-than-truckload) loads the driver might stop at multiple facilities and unload a handful of pallets at each. This would be more likely for smaller-format discount retail like, e.g., the Dollar Store or a restaurant supply chain.

I have family that works for a food distributor where their drivers get to restaurants, unload the truck, stock the items in the fridge/shelves, and take additional orders from the customer before they move on to the next delivery.

Local drivers. They follow a different process than long-haul drivers do and are way more likely to unload their own trucks.

Re: Self-driving trucks begin mail delivery test for U.S. Postal Service

#89

Post by Symbiote »

gpm wrote: post #8
ru999gol wrote: post #3

I always wondered why it seems we've get automated trucks pretty soon everywhere, but there is almost no one talking about self driving trains?

You need a lot more people to move the same amount of cargo in trucks, so there is a lot more money to be saved. The consequences of individual crashes is also higher, which changes the perception and consequences of mistakes, even if the number of trucks means the actual scale of the problem might be similar.

The consequence of crashes is what led to the earliest implementations of automation on railways, starting in the 19th century: https://en.wikipedia.org/wiki/Train_stop

Arguable a mechanical interlocking also counts, as it prevented unsafe routes being set for a train, and made it faster to set the intended route: https://en.wikipedia.org/wiki/Interlocking

Re: Self-driving trucks begin mail delivery test for U.S. Postal Service

#90

Post by esilver »

lm28469 wrote: post #84
esilver wrote: post #51

Earlier quoted context omitted.

The in-vehicle monitor will likely be a standard part of autonomous and semi-autonomous trucking for the foreseeable future. A monitor can be paid less than a driver, as they’re doing less work, and can respond to onsite issues more promptly and accurately than could a remote one.

> A monitor can be paid less than a driver, as they’re doing less work People working full time still need to live you know, even if it's "less work". Last time I checked the "amount" of work doesn't really correlate with the pay too. You can't tell a professional truck driver "Ok man, you'll be in the same seat, on the same road, for the same amount of time but you get $500 instead of $1500 because it's less work, o…

What drivers are paid for a given load reflects market conditions like e.g., weather, truck availability, and load availability, more than anything else.

Automating large parts of the driving process would make the job easier and more desirable. More people would be willing to work as monitors, which would increase truck availability and lower rates.

There’s also the possibility that a monitor could do other work while the truck drove. Team drivers already do this; the passive driver occupies herself when not driving by reading or even dispatching other loads.

Post reply on HN