Waymo Makes History: First Fully Self Driving Car With No Driver

[setoKaiba]

I figured (Google/Waymo) deserves its own thread because they actually so far ahead and the only ones with actual self driving cars although cruise is right in their rear view mirror.

Waymo is first to put fully self-driving cars on US roads without a safety driver

Waymo now testing its self-driving cars on public roads with no one at the wheel

EpsonVoice: How The Latest Big-Screen Tech Can Boost Your Business

Waymo makes history testing on public roads with no one at the wheel


Waymo Safety Report

https://storage.googleapis.com/sdc-prod/v1/safety-report/waymo-safety-report-2017-10.pdf

 

[malcolm]

Nice demo video, but would be more useful to see the three cars tackle very different neighbourhoods / road types.

All three cars seem to be navigating the same burb. Which has its uses, say for a local delivery company or if your name happens to be Truman Burbank.

 

[jaguar36]

How is this any better than what Tesla showed off a year ago?

Canned demo's like this are pretty worthless. Need to be either a much longer video (like a coast to coast one) or it needs to be on a route chosen at random by a third party.

 

[setoKaiba]

How is this any better than what Tesla showed off a year ago?

Canned demo's like this are pretty worthless. Need to be either a much longer video (like a coast to coast one) or it needs to be on a route chosen at random by a third party.

Its not a demo. Nor is it route based.

Waymo, the autonomous vehicle division of Alphabet, Google’s parent company, reached an important milestone recently: since mid-October, the company has been operating its autonomous minivans on public roads in Arizona without a safety driver — or any human at all — behind the wheel. And starting very soon, the company plans to invite regular people for rides in these fully self-driving vehicles.

The news that Waymo’s vehicles have been on public roads with no human in the driver’s seat was announced today by the company’s CEO John Krafcik at a tech conference in Lisbon. The announcement comes on the heels of Waymo’s decision to invite a group of reporters to visit Castle, a 91-acre facility in California’s Central Valley that it has been using as a training course for its self-driving vehicles. At the time, Krafcik declined to provide an exact timetable as to when it would begin testing fully self-driving vehicles on public roads. Little did we know at the time, they were already doing it.

Nice demo video, but would be more useful to see the three cars tackle very different neighbourhoods / road types.

All three cars seem to be navigating the same burb. Which has its uses, say for a local delivery company or if your name happens to be Truman Burbank.

Again, not a demo.

"They’ll be geofenced within a 100-square-mile area of the town of Chandler, a suburb of Phoenix — though Waymo says it plans to expand to areas beyond that as its cars collect more data and conduct more trips."

Here's a blog post from Waymo

Waymo’s fully self-driving vehicles are here – Waymo – Medium

Now, in an area of the Phoenix metro region, a subset of our fleet will operate in fully autonomous mode, with Waymo as the sole driver. Over time, we’ll cover a region that’s larger than the size of Greater London, and we’ll add more vehicles as we grow.

 

[jaguar36]

Again, not a demo.

Yes, yes it is. Guarantee that they ran that exact route multiple times before making the video.

If they were confident in it, they would let journalists give a try instead of just showing a canned demo.


Note, I'm not saying its not further along than Tesla, or that it isn't close to being production ready. Just saying that video is no better than what Tesla showed last year.

Further Note. They give no indication of how much all of their hardware costs. If they need $300k worth of gear on each car to make it work...

 

[malcolm]

It's a demo

Members of the public who live in the geofenced areas can apply to sit in the cars and be taxied around.

From the above link:

Our goal is to give our early riders everyday access to our vehicles

So no one has everyday access at present.

It's a demo. Again, nice demo. But a demo.

 

[Stirthepot]

Its great progress. Competition and progress will benefit all of us. Let's celebrate it!

Thanks OP for sharing

 

[d21mike]

Really happy to see this but you can not say "actually so far ahead" until you or I can actually BUY ONE.... I assume they will have DEMO's like this for multiple years before anyone can BUY ONE. To me TESLA is still far ahead since what they offer is more advance then any other production cars that I can BUY. And as we all know TESLA did a similar DEMO. But until TESLA adds these features to the fleet it is still just R&D.

 

[NerdUno]

Demo obviously means different things to different people. But, if I'm sitting as a passenger in a self-driving car and nobody is in the driver's seat, that is far beyond what I consider a demo to be. The Tesla fanboys are obviously having some difficulty coming to grips with the fact that their savior has officially been left as a tiny spec in Waymo's rear view mirror.

 

[JamesL6]

Really happy to see this but you can not say "actually so far ahead" until you or I can actually BUY ONE.... I assume they will have DEMO's like this for multiple years before anyone can BUY ONE. To me TESLA is still far ahead since what they offer is more advance then any other production cars that I can BUY. And as we all know TESLA did a similar DEMO. But until TESLA adds these features to the fleet it is still just R&D.

I don't think anyone will be able to personally buy one of Waymos cars, there's really no need for them to sell you one. When we get to full self driving, many people (starting in big cities) won't own cars, but just hail them when needed. That's what they are going for

 

[McHoffa]

Demo obviously means different things to different people. But, if I'm sitting as a passenger in a self-driving car and nobody is in the driver's seat, that is far beyond what I consider a demo to be. The Tesla fanboys are obviously having some difficulty coming to grips with the fact that their savior has officially been left as a tiny spec in Waymo's rear view mirror.

but Tesla could, in fact, do a similar demo in a closed area. They could probably train a small fleet to navigate one neighborhood/town (they did this in their video a year ago, driving from one home to the headquarters). I'm glad Waymo did this, because it pushes Tesla, but it's still a limited area demo.

Personally, I'd rather have limited self driving capabilities anywhere in the country than full autonomous in one neighborhood.

Waymo and Tesla are taking different routes to the same destination. Only time will tell which way was the quickest and most painless.

 

[Tam]

...How is this any better than what Tesla showed off a year ago?...

The difference:

Although Tesla shows the demo on its website, its policy remains the same that current real life practice demands a hands-on steering wheel approach (if the system detects enough hands-off, it will penalize driver by turning itself off.)

Waymo has revised its policy: Human engineer(s) are now sitting on the second row leaving the driver seat empty.

It's true that journalists and common folks are not included in the demo but it promises soon!

 

[jaguar36]

The difference:

Although Tesla shows the demo on its website, its policy remains the same that current real life practice demands a hands-on steering wheel approach (if the system detects enough hands-off, it will penalize driver by turning itself off.)

Waymo has revised its policy: Human engineer(s) are now sitting on the second row leaving the driver seat empty.

It's true that journalists and common folks are not included in the demo but it promises soon!

That's only because where Tesla did their demo (CA) they were legally required to have a driver in the seat. Arizona's governor signed an executive order allowing the 'driver' to be located elsewhere. Had Tesla done their demo in Arizona they wouldn't have had anyone in the drivers seat either.

 

[d21mike]

I don't think anyone will be able to personally buy one of Waymos cars, there's really no need for them to sell you one. When we get to full self driving, many people (starting in big cities) won't own cars, but just hail them when needed. That's what they are going for

I believe you are right. Of course with NY a lot of people do not own cars now. They are just using taxi's. This will of course be replaced. I love this technology but at this time I prefer to have my own car. I would just like it to be able to self drive my wife to work and come back home for my use. I just dropped my TESLA off for service and took a UBER home. It was fine and only took 3 minutes for the guy to show up. I was talking to the driver and he has some regular clients that he picks up at 4:45am and drops them off at 5am and that is the start of his day. He said he is pretty busy in the morning. Even here in Southern California you can make a case to not having your own car.

 

[Tam]

...legally required to have a driver in the seat...

I don't think "legal" is a good excuse!

3 different states have allowed cars to roam around without a human driver as long as they obey traffic laws. These are not designed just for testing cars, but they are for driverless cars in general:

9/20/2016 Florida
12/10/2016 Michigan
5/30/2017 Georgia

If Tesla can technologically do what Waymo can do now, it could practice in those three states.

I think Tesla can't because of its technology and not because of lacking of laws for driverless cars in 3 those states.

 

[KarenRei]

The fundamental difference: Waymo uses LIDAR. LIDAR is too expensive to just put on all cars (Waymo has been trying to lower the cost, but it's still too much). Requires a big, ugly, poor-aerodynamics dome on the top. Doesn't work in bad weather (do people not get to drive then?) because it sees rain, snow, blowing debris, etc as obstacles. Etc, etc. Don't get me wrong, when it works, LIDAR gives you an amazingly good data stream, a lot higher quality than photogrammetry (photo stitching to measure distances). But there's a good reason Tesla has deliberately avoided using it.

I'm more hopeful for time-of-flight cameras, which provide a LIDAR-like datastream, from a device the size of a camera (which also returns regular image data, and can be manufactured by similar processes). Tesla could simply swap out their current cameras for time-of-flight ones, and they'd have a new dataset which they can incorporate into their current process, replacing photogrammetry- and radar mapping when weather conditions allow.

 

[AnxietyRanger]

Had Tesla done their demo in Arizona they wouldn't have had anyone in the drivers seat either.

At the stage Tesla's FSD was in October 2016 vs. where Waymo is today, or even was in 2016, I'm pretty sure Tesla more likely needed to have that driver their. Their disengagement reports from the period are much worse than Waymo's.

Waymo has been at this for a long time. It is not surprising they are ahead.

 

[AnxietyRanger]

The fundamental difference: Waymo uses LIDAR. LIDAR is too expensive to just put on all cars (Waymo has been trying to lower the cost, but it's still too much). Requires a big, ugly, poor-aerodynamics dome on the top. Doesn't work in bad weather (do people not get to drive then?) because it sees rain, snow, blowing debris, etc as obstacles. Etc, etc. Don't get me wrong, when it works, LIDAR gives you an amazingly good data stream, a lot higher quality than photogrammetry (photo stitching to measure distances). But there's a good reason Tesla has deliberately avoided using it.

I'm more hopeful for time-of-flight cameras, which provide a LIDAR-like datastream, from a device the size of a camera (which also returns regular image data, and can be manufactured by similar processes). Tesla could simply swap out their current cameras for time-of-flight ones, and they'd have a new dataset which they can incorporate into their current process, replacing photogrammetry- and radar mapping when weather conditions allow.

That's SO, SO misleading.

Waymo uses: 360 degree vision, 360 degree radar and 360 degree lidar. A redundant, 360 degree system.

It most definitely sees in more situations than Tesla does, in more directions.

 

[Tam]

...LIDAR is too expensive...

High LIDAR cost is a problem for average $35,000 consumer cars.

However, current high LIDAR cost is very much affordable for commercial companies such as ride hailing services and trucks.

They are willing to pay a very high price now but they can recuperate the expense by not hiring a driver.

So, I think, technology is still the challenge right now even when commercial companies are willing to pay for it.

 

[RubberToe]

Great post, was going to link to the story before I saw that someone already did.

Where are all the people saying this wouldn't happen for 25 years???

RT