Why AP 2.0 Won't Be Here Soon, and It Won't Be What You Think It Is

[xav-]

I don't mind being wrong. In fact, I'd love to be wrong, in which case I'd be looking for a path to upgrade my car.

But we've only just heard the hen is pregnant, and folks are already counting the chicks. Or something like that.

i agree. All this talks of autonomous vehicle, ride sharing seems so premature to me. i would not consider buying the autonomous software update if I bought the model S today (but would buy EAP)

I would much rather see the focus on addressing the drawbacks of AP today:

- when I see that Chinese kid slamming into a stationary truck at high speed (dieing at the scene)
- when I see the Jerry Brown accident,
- when I see that Swiss dude slamming into a stationary truck at low speed thank god (again all on video)

...I am not sure that the main challenge of AP today is full autonomy.

There are also numerous complaints shared on this forum, such as the car getting too close to center dividers, not handling tight turns, not passing trucks correctly etc. Will AP20 address that?

 

[3Victoria]

Those accidents were all due to documented limitations of the early hw/sw. The system ignored stationary returns from the radar, since there are many such returns, such as posts, concrete blocks, road surface changes etc which would have caused braking if not ignored. This was documented specifically in the cars' manuals. In addition, the drivers were and are responsible for the driving. These accidents were caused by people, not AP. One can argue about 'loss of attention' and we have had those repeatedly. Tesla, who knows more about their requirements, thinks AP2 HW is sufficient. While they may be wrong, their opinion is much better than the peanut gallery, and I include myself in that category.

 

[HankLloydRight]

d. This was documented specifically in the cars' manuals

This is the crux of the problem. Nobody RTFM. When the next level of Tesla AP hits, A LOT of people are going to assume it's fully autonomous, or close (Model 3 owners?), and are NOT going to be paying the attention they should be paying to actually, you know, DRIVING.

This is going to lead to many more serious/fatal accidents inappropriately blamed on AP (because the public and press are dumb), which will set back the legal and liability issues even further.

I just don't see a world with driverless UBERs ("Johnny Cabs") driving people around in at least 10 years if not longer. But some people think that's coming in 2017 or 2018.

 

[Joe F]

RE camera placement mentioned earlier (sorry, just getting caught up), how many here have had to hug the left side of the lane, and in some cases, shift their head left as well, in order to see around a larger vehicle blocking the view of traffic ahead?

While increasing following distance is called for in this case, it's not enough to see what traffic ahead is doing, IMHO.

While I understand the camera placement form a wiper cleared view perspective, how useful are the views of the back of a truck?

Off highway with traffic lights, following a tractor that runs a changing stoplight, which the cameras can't possibly see, what happens then?

Imagine driving from the middle of your car. Feel comfortable?

 

[u00mem9]

i agree. All this talks of autonomous vehicle, ride sharing seems so premature to me. i would not consider buying the autonomous software update if I bought the model S today (but would buy EAP)

Exactly. It's like a common shady sales technique to start elaborating on all the implications and benefits of a bold claim, rather than detailed statements about what is actually confirmed and proven.

Doesn't bode well.

 

[calisnow]

This is the crux of the problem. Nobody RTFM. When the next level of Tesla AP hits, A LOT of people are going to assume it's fully autonomous, or close (Model 3 owners?), and are NOT going to be paying the attention they should be paying to actually, you know, DRIVING.

This is going to lead to many more serious/fatal accidents inappropriately blamed on AP (because the public and press are dumb), which will set back the legal and liability issues even further.
.

Everything will be fine. AP 1.0 is fine even though the same was said of it. One (or possibly two) deaths in over 200 million miles of driving. A number of accidents later and it is still turned on in cars, still called Autopilot, Tesla's sales are still climbing. Everything is fine. And this is the most primitive autopilot that will ever exist.

 

[David29]

I suppose you saw what I saw TODAY on the news. An OTTO Autonomous Equipped truck drove from warehouse to warehouse with a semi load of beer, with no driver in the driver's seat (he was watching from the cab sleeping quarters) for the entire 150 miles. Why can Tesla not be able to do this and better, in even another month, if not even already doing it?

Anything you or I think is pure speculation.

Yeah, I thought that was pretty impressive, especially in the sense that the demo involved no one in the front 2 seats. So any intervention would have taken some time for the human to move into the driver's seat. Shows considerable confidence.

The article I read (in Wired, i think) mentioned that the hardware cost $30K. If so, that is presumably way beyond what is being spent on Tesla's hardware. But hardware aside, i do wonder about the apparently quite advanced state of the software. E.g., to use a scenario postulated a few pages above, would the truck's software have been able to deal with an unexpected interruption to traffic flow such as an accident, with a cop or a flagman directly traffic? Can you program the system to recognize a human making vague hand gestures or waving a flag, to indicate to "stop" or "go that way?"

And if the company that built the truck control system does have that knocked, wouldn't it make sense for Tesla or any other manufacturer to buy the solution rather than to start from the proverbial clean sheet of paper?

Unless that truck demo was somehow less than it appeared, I thought it was rather compelling demonstration....

 

[MP3Mike]

Yeah, I thought that was pretty impressive, especially in the sense that the demo involved no one in the front 2 seats. So any intervention would have taken some time for the human to move into the driver's seat. Shows considerable confidence.

Unless someone had a remote control connection to take over instantly if necessary.

 

[HankLloydRight]

I wonder if there was a lead car 200 feet in front of the truck that never made it into a clear shot.

 

[MP3Mike]

And maybe following cars preventing passing, etc.

 

[HankLloydRight]

I wonder what kind of state and local clearances they needed for that to make it legal.

 

[maxell2hd]

A number of articles mentioned the truck had a police escort, and the trip began at 1AM so traffic would have been light.

 

[MP3Mike]

A number of articles mentioned the truck had a police escort, and the trip began at 1AM so traffic would have been light.

It was way too light out for it to have been shot at 1 AM.

 

[ohmman]

It was way too light out for it to have been shot at 1 AM.

They have aggressive daylight savings time there.

 

[MP3Mike]

OK, I found some information in this article: "Uber's Otto team worked with Colorado regulators to get permission for the delivery and to arrange for police supervision of the shipment, said Ron. Otto spent two weeks scoping out the driving route from Fort Collins to Colorado Springs, carefully mapping the road to make sure the technology could handle it. The team wanted the trip to take place in the early morning when traffic would be relatively light and on a day when the weather was clear. Those conditions were met last Thursday, when the delivery took place."

 

[maxell2hd]

It was way too light out for it to have been shot at 1 AM.

I think all those shots were promo shots and not from the actual trip. A couple of articles below place the time at 12:30AM to 1:00AM when they left Fort Collins:

http://uk.businessinsider.com/uber-otto-self-driving-truck-completes-first-run-2016-10
Uber’s Self-Driving Truck Makes Its First Delivery: 50,000 Budweisers

 

[HankLloydRight]

So in other words, almost completely staged. Not surprised.

 

[3Victoria]

There have been lots of demos over known and limited routes. This demo is not too impressive, but promoted well.

 

[bob_p]

How was the truck demo different than AP 1.0 - on ramp to off ramp driver assist?

 

[Joe F]

How was the truck demo different than AP 1.0 - on ramp to off ramp driver assist?

No hands hovering near steering wheel?
Big ass box with a few more than 4 wheels?
Few more gears to shift?
Slight risk of jackknifing?
Bigger horn?
Spilling crappy Bud all over the highway?