Waymo

[Daniel in SD]

Yes - kind of weird.

No wonder they are suing.

Isn't it interesting that on a Tesla dedicated website some people are always demanding more transparency from Tesla and less from others ? I wonder what the motivations are.

The DMV also redacted (or at least tried to redact ) some communications with Tesla too.
All people are demanding from Tesla is that they comply with AV testing reporting requirements.
Also it's not surprising that people are more interested in Tesla on a Tesla forum. I don't think anyone is demanding less transparency from Waymo...

 

[diplomat33]

Isn't it interesting that on a Tesla dedicated website some people are always demanding more transparency from Tesla and less from others ? I wonder what the motivations are.

Nobody is demanding more transparency from Tesla and less transparency from others. I think most people want equal transparency. The fact is that Waymo is fully transparent with the CA DMV and complying with all of the CA DMV regulations. Waymo just wants some parts redacted from the public report. But the CA DMV gets to see everything. Tesla is not transparent with the CA DMV at all. I think people just want Tesla to be as transparent with the CA DMV as everyone else is.

 

[powertoold]

We all know how this is gonna end

 

[Daniel in SD]

We all know how this is gonna end

How?
We have no idea if what was redacted is legitimately a trade secret or not. I don't even know what the Freedom Of Information Act says about trade secrets! Unless we have a lawyer here who's an expert on FOIA requests I would say we have no idea.

 

[Daniel in SD]

Moving this to the Waymo thread!

Companies have forever prevented people from knowing what is going on based on "trade secrets" BS.

For eg. companies won't disclose what chemicals they are pumping into get fracking gas out - claiming "trade secrets". When public interest trumps trade secrets, they should be disclosed.

Well one could argue that we know how dangerous Waymo's testing program is because they disclose all collisions. Is knowing more details about their system necessary for public safety? I have no idea...

Here is the redacted Waymo correspondence with the DMV "Re: DMV Deployment Follow-Up"

https://ca-times.brightspotcdn.com/46/3c/bbe1b7b4473da941a00851c82852/redactions-part-215.pdf

 

[Daniel in SD]

Moving this to the Waymo thread!

Well one could argue that we know how dangerous Waymo's testing program is because they disclose all collisions. Is knowing more details about their system necessary for public safety? I have no idea...

Here is the redacted Waymo correspondence with the DMV "Re: DMV Deployment Follow-Up"

https://ca-times.brightspotcdn.com/46/3c/bbe1b7b4473da941a00851c82852/redactions-part-215.pdf

I have to say the redacting of some of the DMVs questions seems odd. I guess they could be referencing previously disclosed “trade secrets”.

 

[EVNow]

Moving this to the Waymo thread!

Well one could argue that we know how dangerous Waymo's testing program is because they disclose all collisions. Is knowing more details about their system necessary for public safety? I have no idea...

Here is the redacted Waymo correspondence with the DMV "Re: DMV Deployment Follow-Up"

https://ca-times.brightspotcdn.com/46/3c/bbe1b7b4473da941a00851c82852/redactions-part-215.pdf

Thanks for the link. The redactions look very, very extensive. Many times whole pages. This is a mockery of the FOIA.

ps : I've no special insights into Waymo case in particular - but in general our judicial system has been pro-company secrets and anti-public disclosure for decades now. That always shapes what companies think they can get away with.

For eg. look at page 14

 

[diplomat33]

Waymo won its FOIA lawsuit to keep some details secret:

Waymo to keep robotaxi safety details secret, court rules – TechCrunch - dramaticonline.com

Waymo, the autonomous driving arm of Alphabet, was granted a win on Tuesday when a California court ruled it could keep certain details regarding its AV

dramaticonline.com

 

[willow_hiller]

Waymo won its FOIA lawsuit to keep some details secret:

Waymo to keep robotaxi safety details secret, court rules – TechCrunch - dramaticonline.com

Waymo, the autonomous driving arm of Alphabet, was granted a win on Tuesday when a California court ruled it could keep certain details regarding its AV

dramaticonline.com

I'm guessing courts are going to initially defer to company claims when it comes to information disclosure, at least until the case is appealed. Because you cannot un-disclose information, once that genie is out of the bottle.

 

[diplomat33]

I posted this in the autonomous progress thread but this thread is probably more appropriate.

Waymo posted a new blog explaining how they use "key points" to better predict behavior of road users. Put simply, instead of using bounding boxes, camera vision detects points on the human body (like ankle, knee, shoulder, hands, head etc) and connects those key points to get a representation of the body pose and gestures. This methods allows for more detailed prediction of behavior. In the blog, Waymo has some clips of their perception where you can see the car detecting lots of pedestrians, represented as "stick figures". Some example of how Waymo uses "key points": predicting behavior of large crowds, detect gestures of human traffic controllers or cyclists, detect indecisive pedestrians crossing the street, and detect partially occluded objects like detecting a pedestrian from just a leg sticking out from behind a car.

Key points are a compact and structured way to convey human pose information otherwise encoded in the pixels and lidar scans for pedestrian actions. These points help the Waymo Driver gain a deeper understanding of an individual's actions and intentions, like whether they’re planning to cross the street. For example, a person's head direction often indicates where they plan to go, whereas a person's body orientation tells you which direction they are already heading. While the Waymo Driver can recognize a human's behavior without using key points directly using camera and lidar data, pose estimation also teaches the Waymo Driver to understand different patterns, like a person propelling a wheelchair, and correlate them to a predictable future action versus a specific object, such as the wheelchair itself.

The Waymo Driver uses real-time data from our sensor suite, including our lidars, which feed into our neural-network models to localize key points in three dimensional space. Waymo created its own methodologies to generate high-quality labels to identify the joints in a 3D space, which enabled training human pose models to further improve the safety of the Waymo Driver. This also means that Waymo's key point technology doesn’t identify an individual person, but rather aggregates data points and provides us with a better capability to recognize that a person exists and where they may be going, which is especially beneficial for partially visible pedestrians that might be stepping out of a vehicle or sitting near the road. Additionally, we’ve optimized our system to run onboard the vehicle in real-time, with high precision and low latency, to enhance its behavior-prediction models and allow the Waymo Driver to quickly and safely handle any situation.

Utilizing key point and pose estimation for the task of autonomous driving

Historically, computer vision relies on rigid bounding boxes to locate and classify objects within a scene; however, one of the limiting factors in detection, tracking, and action recognition of vulnerable road users, such as pedestrians and cyclists, is the lack of precise human pose...

blog.waymo.com

I find the work really interesting. IMO, it is an example of some of the advanced capabilities of Waymo's autonomous driving because their perception is going beyond just detecting objects and actually able to detect movement of points on the body for hundreds of users to better predict behavior. @Bladerskb

 

[EVNow]

I find the work really interesting. IMO, it is an example of some of the advanced capabilities of Waymo's autonomous driving because their perception is going beyond just detecting objects and actually able to detect movement of points on the body for hundreds of users to better predict behavior.

Reminds me of Kinect.

 

[diplomat33]

Reminds me of Kinect.

I think Kinect probably used a similar concept, detecting "key points" of the players and translating their actions into a game. "Key points" is not a new technology. What Waymo is doing is applying the technology to the problem of autonomous driving which is orders of magnitude more challenging than how "key points" have been used before:

While we incorporate this exciting technology onboard our vehicles to help the Waymo Driver navigate the real world, key points and pose estimation have been advancing many industries for more than a decade –- from animating beloved cartoons and creating realistic video game characters to augmenting reality on popular social media apps – but applying this state of the art technology to the autonomous driving domain is magnitudes more challenging.

Up until now, key points have been used in relatively controlled environments to help make them easier to apply, such as augmenting a dinosaur next to a singular person or filming a set number of actors to control a video game. The Waymo Driver generates key points in the "wild" for all nearby road users, which is orders of magnitudes harder as our Driver often encounters up to hundreds of pedestrians at a single intersection, many of which can be occluded by other objects.

 

[EVNow]

I think Kinect probably used a similar concept, detecting "key points" of the players and translating their actions into a game. "Key points" is not a new technology. What Waymo is doing is applying the technology to the problem of autonomous driving which is orders of magnitude more challenging than how "key points" have been used before:

I guess animation was probably the first industry to use it - obviously the one with most benefits. But I think they use multiple sensors instead of camera only.

May be the tech has now advanced to capture facial expressions using cameras too ....

 

[diplomat33]

Nice edge case where cameras alone would be blinded at night by the incoming headlights and would have trouble seeing the pedestrian. Thanks to sensor fusion with lidar, Waymo handles the situation very safely.

 

[ThomasD]

Got to make sure they stop for those Gators in the crosswalks.

 

[diplomat33]

Watch those Gator crosswalks

Why are you posting this? What's your point?

 

[ThomasD]

Can the cars be programed to handle unpredictable events Like a cop on a Horse directing traffic.

 

[diplomat33]

Can the cars be programed to handle unpredictable events Like a cop on a Horse directing traffic.

Yes.

 

[EVNow]

Got to make sure they stop for those Gators in the crosswalks.

A lot more well behaved than Florida drivers for sure

 

[DanCar]

Can the cars be programed to handle unpredictable events Like a cop on a Horse directing traffic.

Can they? Yes. Will they? Who knows. Time will tell. I'm sure there will be events they won't handle for the next ten years.