-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Waymo
- Thread starter Daniel in SD
- Start date
diplomat33
Average guy who loves autonomous vehicles
The image looks blurred so maybe that's part of the issue. Who knows, the blurring could be from overlaying multiple scans.
There could also be different lidar systems for distant and near FOV use. If so the distant scanning lidar could use a circular scan whereas the other, possibly near FOV, uses a more random scan pattern. On the other hand the near FOV return surfaces could be uneven making it look that way. But I agree, the scan patterns change in that image.
powertoold
Active Member
The best explanation was from nvx2777 (sp.). It seems Waymo purposely increased the size of the point cloud dots from the roof LIDAR to emphasize it for our consumption. The roof LIDAR dots seem to be ~8x larger than the other dots.
diplomat33
Average guy who loves autonomous vehicles
Brad Templeton's Waymo driveless ride with discussion with a senior developper at Waymo:
0:00 Intro
1:14 Begin ride
2:12 Nudge around truck
2:22 What are the blockers to deployment?
3:53 Gap in video explaned
4:03 Onto Geary
4:20 Machine learning in planner
5:50 TPUs in vehicle
6:30 nudge out
7:50 Sim
8:20 About unprotected lefts
9:34 USF stop, not really pulling over
12:50 Competition in some cities
15:04 on Tesla
16:00 Not just a video
16:50 Honks and detecting them
20:00 Going around a double parked car, then quickly finding lane again rather than human style
21:10 We finish, stopping in front of bike parking and blocking lane a bit
0:00 Intro
1:14 Begin ride
2:12 Nudge around truck
2:22 What are the blockers to deployment?
3:53 Gap in video explaned
4:03 Onto Geary
4:20 Machine learning in planner
5:50 TPUs in vehicle
6:30 nudge out
7:50 Sim
8:20 About unprotected lefts
9:34 USF stop, not really pulling over
12:50 Competition in some cities
15:04 on Tesla
16:00 Not just a video
16:50 Honks and detecting them
20:00 Going around a double parked car, then quickly finding lane again rather than human style
21:10 We finish, stopping in front of bike parking and blocking lane a bit
Last edited:
diplomat33
Average guy who loves autonomous vehicles
Interesting review from someone took nearly 50 public driverless rides (not trusted tester) in downtown Phoenix:
My impression is that the Waymo self-driving tech is very good. But Waymo is not there yet as a replacement for personal car ownership. Waymo needs to work on lowering cost, expanding the service areas, improving convenience of pick up and drop offs and adding more cars to lower wait times. Up to now, Waymo has focused almost entirely on just achieving good autonomous driving. This made sense since they needed to build safe and reliable autonomous driving first. But now that they have really good autonomous driving, I suspect that Waymo will probably focus more on these operational issues.
My impression is that the Waymo self-driving tech is very good. But Waymo is not there yet as a replacement for personal car ownership. Waymo needs to work on lowering cost, expanding the service areas, improving convenience of pick up and drop offs and adding more cars to lower wait times. Up to now, Waymo has focused almost entirely on just achieving good autonomous driving. This made sense since they needed to build safe and reliable autonomous driving first. But now that they have really good autonomous driving, I suspect that Waymo will probably focus more on these operational issues.
Last edited:
jabloomf1230
Minister of Silly Walks
In case anyone was wondering how the recently announced Google layoffs affects Waymo, there doesn't appear to have been any recent layoffs at Waymo:
Waymo appears to be hiring:
Waymo appears to be hiring:
jabloomf1230
Minister of Silly Walks
Which leads me to ask. Has anyone developed a credible financial analysis of robotaxis vs. taxis and ride-sharing?
Doggydogworld
Active Member
I've seen everything from 14 cents/mile and complete domination of total VMT (Seba, etc) to $2++/mile limited to a few city centers. Everyone is just guessing, including the serious-looking "market analysts" with seriously-priced reports. Waymo and Cruise plus a few Chinese companies have some real-world data on costs, but their services are too limited to lend much insight into consumer acceptance.
Hopefully Waymo is advertising for Superbowl fans to use the service to/from the Phoenix airport.
diplomat33
Average guy who loves autonomous vehicles
I think it is too early to
Yes they are. They have a sign at the airport.
They also tweeted about it:
Yes they are. They have a sign at the airport.
They also tweeted about it:
All I know is that Uber is about $2 a mile here. That would be a lot of revenue over the life of a robotaxi.
For wide-scale deployment I expect to see some type of dynamic demand control pricing similar to airlines, trains, hotels, and other capital intensive entities.
That last-minute peak-period commute to work is going to be expensive, but you might pay 15 cents/mile for booking a recurring trip every Saturday at 3pm for 12 weeks to baseball practice. A bartender heading home at 4am might even find vehicles for 10 cents/mile, just enough to cover the operating cost and allow some marketing (see various $1 flight and inter-city bus marketing).
This will happen fairly naturally once multiple companies with high fixed costs and low marginal costs start competing for market share. If a charge of $1.01 contributes 1 cent toward fixed costs, then they'll chase that customer if their vehicle would otherwise be idle.
Last edited:
You may have different rates after a certain mileage. The first mile maybe 15 cents then maybe an extra 75 cents after that. The farther the car drives at .15 cents a mile the more money will be lost for the company. Your cleaning/Maintenance/charging costs will exceed what the car will make per day at such a low rate.
Doggydogworld
Active Member
Per mile with surge pricing is fine for taxi/Uber replacement, but I don't think it will fly in the much larger personal car replacement market.
I expect monthly subscriptions, like the old cell phone plans with 300/500/unlimited peak minutes. It's much easier to sell a quasi-fixed cost that people can compare directly to their existing monthly car payment + insurance + parking cost. Maybe a tier with unlimited peak carpool miles or something, too. And throw in some Zipcar credits for the occasional errand outside the geofence.
Similar threads
- Article
