Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register

Waymo

This site may earn commission on affiliate links.
From a Waymo employee.

So just an anonymous source?

Does Waymo tell you they cancelled long haul testing between Arizona and Texas?

Why does this even matter? Waymo is moving forward with commercial autonomous trucking in TX. That's is better than testing. This shows that their testing between AZ and TX paid off and the Waymo Driver can now handle L4 highway.
 
  • Like
Reactions: Microterf
So just an anonymous source?
Yes, anonymous to the internet, not anonymous to me.

Why does this even matter?
General trend I've seen with Waymo. Try to over market themselves with good news and not state bad news. Perhaps true of any startup. Also true with Tesla / Elon.

Waymo is moving forward with commercial autonomous trucking in TX.
Is there a safety driver?

That's is better than testing. This shows that their testing between AZ and TX paid off and the Waymo Driver can now handle L4 highway.
I remember, seems like two years ago, where Waymo said they don't bother testing on highways because it is too easy. It is still just testing. If there were no safety driver then that would be more impressive.
 
Last edited:
  • Like
Reactions: mikes_fsd
Yes, anonymous to the internet, not anonymous to me.


General trend I've seen with Waymo. Try to over market themselves with good news and not state bad news. Perhaps true of any startup. Also true with Tesla / Elon.


Is there a safety driver?


I remember, seems like two years ago, where Waymo said they don't bother testing on highways because it is too easy. It is still just testing. If there were no safety driver then that would be more impressive.
They are testing with no backup driver? What happens when the truck gets off the highway? I didn't know it could handle side and back roads. The trucks can back into loading docks? Do they have remote drivers for emergencies such as a detour. Where the truck may have to navigate an unmapped area
 
They are testing with no backup driver?

We don't know if Waymo will use safety drivers or not for these deliveries.

What happens when the truck gets off the highway? I didn't know it could handle side and back roads.

It's the same Waymo Driver that they use in Chandler on city streets. So we know the Waymo trucks can handle city streets. Presumably the truck could continue to drive autonomously off the highway.

The trucks can back into loading docks?

We don't know.

Do they have remote drivers for emergencies such as a detour.

Waymo uses remote operators for the ride-hailing in Chandler. So I think we can assume that the trucks would either have remote operators or safety drivers.

Where the truck may have to navigate an unmapped area

Waymo does not drive outside mapped areas. And the advantage of deliveries is that the trucks would follow the same route every time since they only drive between where they load up and where they deliver.
 
  • Like
Reactions: Microterf
It seems like Google/Waymo is driving around Mountain View again.

I've been seeing unmarked white cars driving around Mountain View for a while now. They had a lot of sensors, including a big lidar/camera tower in the middle of the hood like for their Mazda 3 Maps and Street View cars. I had an inkling they were Google/Waymo, but didn't know. I finally spotted a Jaguar iPace (white, unmarked) with lidars so I think that's more or less confirmation that it's Google/Waymo.

I haven't seen any Pacificas in a long time.
 
I suspect that Waymo will slowly fade out the Pacificas since they have the older generation hardware and transition fully to the new 5th generation hardware on the Jaguar i-Pace.
They can put Gen5 sensors on anything (not that Gen5 makes much difference, anyway). The i-Pace is not a good Robotaxi platform. Fewer, less comfortable seats with less storage and poorer ingress/egress. It's also expensive, with the money being wasted on stuff like performance which Waymo doesn't use.

I'm certain they have a custom pod in the works with Magna. If they were going to scale up soon they'd use Pacificas, IMHO. But they won't scale, so it's moot.
 
They can put Gen5 sensors on anything (not that Gen5 makes much difference, anyway). The i-Pace is not a good Robotaxi platform. Fewer, less comfortable seats with less storage and poorer ingress/egress. It's also expensive, with the money being wasted on stuff like performance which Waymo doesn't use.

I'm certain they have a custom pod in the works with Magna. If they were going to scale up soon they'd use Pacificas, IMHO. But they won't scale, so it's moot.

Yes, they could retrofit Pacificas with the 5th Gen but I have not seen any yet. As far as I know, Waymo has only put the 5th Gen on the I-Pace so far.

And it is not moote because Waymo will absolutely scale.
 
New blog on how Waymo has developed the Waymo Driver to safely handle cyclists:


Clip of 5th Gen I-Pace and cyclists in SF:


The lidar point cloud of the 5th Gen is very accurate and detailed. That is some amazingly accurate perception. That blows my mind!
 
Last edited:
New blog on how Waymo has developed the Waymo Driver to safely handle cyclists:


Clip of 5th Gen I-Pace and cyclists in SF:


The lidar point cloud of the 5th Gen is very accurate and detailed. That is some amazingly accurate perception. That blows my mind!

incredible detail!
 
  • Like
Reactions: diplomat33
New blog on how Waymo has developed the Waymo Driver to safely handle cyclists:


Clip of 5th Gen I-Pace and cyclists in SF:


The lidar point cloud of the 5th Gen is very accurate and detailed. That is some amazingly accurate perception. That blows my mind!
Waymo's technology always impresses me. Resolution may be higher than AEye (1550 nm, MEMS). It's interesting how they subtract out the buildings and other permanent features (well, almost, parts of some trees make it in). I wonder if this is just for the visualization, or does it more deeply reflect how they interpret the scene?

I'm really curious why the parked Vespa at 0:40 doesn't show on the point cloud. Same with the motorcycle parked one more car up the hill. Switching to the video feed, what the heck is happening with the colors on left side in the last 5-10 seconds?

The 4 way stop at 0:45 is interesting. Waymo and the white Model X stop at almost exactly the same instant, then both start forward. The Model X should have the right of way, since it is to the right, but it yields and the Waymo proceeds. And Waymo doesn't even flinch when it's less than halfway through the intersection and the Model X starts forward again. It all comes off seamlessly, but I wonder how much is just dumb luck.
 
  • Like
Reactions: diplomat33
Waymo's technology always impresses me. Resolution may be higher than AEye (1550 nm, MEMS). It's interesting how they subtract out the buildings and other permanent features (well, almost, parts of some trees make it in). I wonder if this is just for the visualization, or does it more deeply reflect how they interpret the scene?

Not sure. But it would make sense if it was part of how they interpret the scene since static buildings don't need to be visualized. And it would make sense to reduce noise or clutter, since you want the car to focus on what really matters. Furthermore, substracting out the clutter would reduce the amount of data that the computer has to process. There is enough perception data for the computer to worry about on the road, you don't want to clutter things up with a lot of useless data.

I'm really curious why the parked Vespa at 0:40 doesn't show on the point cloud. Same with the motorcycle parked one more car up the hill. Switching to the video feed, what the heck is happening with the colors on left side in the last 5-10 seconds?

Perhaps, because it was stopped with nobody on it, it was substracted out. It was not in the path of the car and it was not going to move. So it was not an object that the Waymo needed to worry about?

The 4 way stop at 0:45 is interesting. Waymo and the white Model X stop at almost exactly the same instant, then both start forward. The Model X should have the right of way, since it is to the right, but it yields and the Waymo proceeds. And Waymo doesn't even flinch when it's less than halfway through the intersection and the Model X starts forward again. It all comes off seamlessly, but I wonder how much is just dumb luck.

Could be dumb luck. Again, we just have one data point. Although, I know Waymo is focused on driving policy. I would not be surprised if the 5th Gen Driver actually has better driving policy for situations like this. It is definitely a case that an autonomous car needs to handle.
 
Last edited:
The 4 way stop at 0:45 is interesting. Waymo and the white Model X stop at almost exactly the same instant, then both start forward. The Model X should have the right of way, since it is to the right, but it yields and the Waymo proceeds. And Waymo doesn't even flinch when it's less than halfway through the intersection and the Model X starts forward again. It all comes off seamlessly, but I wonder how much is just dumb luck.
Watching the video back 3 or 4 times, its extremely hard to tell, but it looks like the Waymo car stops just a tiny bit before. A human following the rules would probably let the car on the right proceed first, but I'm wondering if the Waymo car is calculating it as "I stopped 42ms before them, I go first."
 
  • Like
Reactions: diplomat33
Perhaps, because it was stopped with nobody on it, it was substracted out. It was not in the path of the car and it was not going to move. So it was not an object that the Waymo needed to worry about?
But they show every parked car in fine detail, even though they are not in Waymo's path and not moving.

A parked car could contain an unseen driver, of course, who might suddenly pull out or open the door and step out into traffic. The Vespa clearly has no rider and hides no such potential threats. The motorcycle could have a rider crouched behind it, though.

I'm probably reading way too much into the visualization, especially since it may have been post-processed.
 
  • Like
Reactions: diplomat33
But they show every parked car in fine detail, even though they are not in Waymo's path and not moving.

A parked car could contain an unseen driver, of course, who might suddenly pull out or open the door and step out into traffic. The Vespa clearly has no rider and hides no such potential threats. The motorcycle could have a rider crouched behind it, though.

I'm probably reading way too much into the visualization, especially since it may have been post-processed.
The poles in the road at 0:17 could be a significant feature but aren't shown.
 
It seems like Google/Waymo is driving around Mountain View again.

I've been seeing unmarked white cars driving around Mountain View for a while now. They had a lot of sensors, including a big lidar/camera tower in the middle of the hood like for their Mazda 3 Maps and Street View cars. I had an inkling they were Google/Waymo, but didn't know. I finally spotted a Jaguar iPace (white, unmarked) with lidars so I think that's more or less confirmation that it's Google/Waymo.

I haven't seen any Pacificas in a long time.
All the I-PACEs I've seen in Arizona say "Waymo Engineering" on the side in a somewhat out of the way place, is there at least that?
 
  • Like
Reactions: diplomat33