Autonomous truck cleared to drive on US roads for the first time

[Recursion]

http://www.newscientist.com/article/dn27485-autonomous-truck-cleared-to-drive-on-us-roads-for-the-first-time.html

The next big thing in autonomous vehicles really is big. At a ceremony at the Hoover Dam last Wednesday, automotive manufacturer Daimler unveiled a self-driving truck – the first to be cleared to drive on US roads.

For the freight industry, the Inspiration Truck holds the promise of a future with fewer accidents, lower fuel costs and well-rested drivers.

Over the past few years, autonomous trucks have drawn the attention of companies that repeatedly use the same routes or encounter few people or other vehicles. Some farms use autonomous grain harvesters or planters. Mining company Rio Tinto has more than 50 self-driving vehicles hauling iron ore at a remote site in Western Australia. In Texas, the US military has been working on trucks that can navigate battle zones.

The Inspiration is different, designed to travel on the highway alongside regular cars and trucks. With clearance to drive on Nevada's highways, this could be big news for the trucking industry, which struggles to find drivers to do the exhausting work. If successful, other big self-driving vehicles could follow, such as garbage trucks or city buses.

[NYC_SKP]

1. I know people, idiots, who say that V2V smart car "trains" are just a couple years away.

I know that models are coming out soon that will communicate with other cars, part of a safety program, and that futurists have talked about these car trains for years.

This same idiot thinks we can build our way out of our overpopulation problem and water shortage, just build desal plants!

He hates rail, loves smart cars.

Fucking idiot.

These trucks, however, are smart trucks not likely to eliminate drivers altogether and I'll be we won't live to see truck on our public highways with no drivers whatsoever, but we'll see!

[Recursion]

2. My buddy at Google privately predicts 2020

These trucks, however, are smart trucks not likely to eliminate drivers altogether

True; I was just stick with the 4-para limit.

I'll be we won't live to see truck on our public highways with no drivers whatsoever

Maybe not, but that's a shame, because humans are very bad at driving vehicles.

Anyways, my buddy at Google very seriously thinks you'll start seeing driverless cars on the road in 5 years, and people who still drive their own cars will be jeered at as Luddite cranks (and pay obscenely high insurance rates) by 2030. And people won't own cars anymore; you'll rent them per-trip.

[NYC_SKP]

4. I've seen one driverless Google car on 580. I expect to see a few more.

But this idiot thinks that the trouble on the PCH, deadly dead stopped traffic, will be solved with V2V.

He fails to realize that congestion on surface streets currently and will continue to back things up on the offramps and onto the highway and it will continue to look like a parking lot.

Anyway, he's anti-rail and V2V will solve all our problems.

[Recursion]

6. Robot cars can cooperate in a way human drivers can't (or at least don't)

Traffic would be a lot better if people would just merge correctly. It's a utopian dream, I know, but robot cars could actually be programmed to do it, which would mean that the merge delays that result from everybody trying to get just a little bit more than the next guy would go away.

(Now, no doubt somebody will sell car firmware hacks that get you the extra little bit more than the other guy, but maybe that could be made illegal or something...)

[NYC_SKP]

7. The person who can't tell if they're sure thay want to go into the mall driveway or not...

...blocks the whole thing.

The system works well only when all cars are using it and all drivers are doing it right.

I'm all for it, I'm just not going to make projections about it solving problems that, in the end, are human problems.

Good point about merging and gaming the software.

The market now trades at light speed and if a two trades are made at the same instant, and one trader is one block closer to the stock market mainframe, there's an advantage.

http://www.radiolab.org/story/267195-million-dollar-microsecond/

[NBachers]

3. How do they work in the snow, where the lane markings are obscured?

How do they react to the idiot in the four-wheeler that cuts in front and jams on their brakes?
How will they respond in a fog or snow white-out?
Do they work in crowded stop-and-go traffic?

[Recursion]

5. In all cases, the answer is "better than humans"

For the snow and fog, the smart roads have IR or FM beacons the cars use to guide themselves, and beacons on the cars to avoid collision. And even better, a computer can reliably read a risk matrix and decide the conditions are too hazardous to continue (humans are very bad at that).

In the crowded stop and go traffic, they actually maintain the distance:speed ratio humans are taught to maintain but never have the patience to.

[denbot]

8. Baugh!

I'll worry about a logarithm doing a better job at driving a big rig then me, when we can maintain a major roadway without potholes. Now, imagine how likely even a weak republican caucus would allow voting for the funding to create, say, a Highway I-80 with embedded IR/FM, and a network with the bandwidth required to communicate a myriad of data to each node; ie; vehicle.

Now imagine the investment required to extend that network to every conceivable delivery point.

Has for a computer to reliably "read a risk matrix", would you bet your life on a technology as well developed as a GPS unit? I use them on a regular basis, the best available for commercial drivers. I don't think any sane person would completely trust their GPS, a well developed, and relatively simple technology.

Baugh!

[Recursion]

9. Human drivers' safety records are pretty discouraging (nt)

Last edited Mon May 11, 2015, 04:46 AM - Edit history (1)

[denbot]

10. Compared to..?

I just don't see a 80,000lbs rig run point to point safely by wire. I run a new rig but even with a new truck, sensors can, and too often, do fail. What happens when a critical inertia sensor, or worst, a communication node fails?

Billions have been spent developing fly by wire for the U.S. Military. After 30 years, with nearly limitless funds, these systems are critical to fly complicated aircraft like the F-22, and those systems fail.

What happens when a far cheaper built truck system fails in poor conditions on a crowded section of road?

The sky, like the ocean is relatively empty. Not so for our very crowded roads and highways.

[Recursion]

11. I think rail is the only medium that's been pretty heavily automated

But those do pretty well. And, yeah, trucks have an additional degree of freedom that trains don't.

I just meant that human drivers are unsafe enough that it would take some argument to convince me that robots would be less safe.

[logosoco]

12. Who is going to blow the air horn when little kids pump their fists at them?

Seriously though, this is an interesting concept. I enjoy driving very much, but I think that is because I mainly travel in low populated areas.
I drove through the desert this past year and I could see how doing that on a weekly basis would really be hard on a truck driver.