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

Tesla Autopilot maps

This site may earn commission on affiliate links.
Probably the same database, but the AP1 cars read speed limits signs as the primary method and only fall back to the database when it can't. AP2 cars currently only use the database. (I would have hoped that they use the AP1 reading signs to improve the database, but it doesn't seem like they do.)

It’s not there database to update it’s Tom tom’s, so I bet there is a process to update them, that takes far to long and entirely to much validation so that it’s clearly outdated information before it gets released.
 
  • Like
Reactions: croman
Amazing, sir. So the branches are called dev, live, stage.

If I look at stage, there's a veeeeery curious map tile name:
na20180212_localizer_optimal: 2018-02-12 15:12



It's newer than those December-ish live tiles we saw, and "localizer_optimal" sounds interesting too :D


EDIT: Paging @DamianXVI , @verygreen , @BigD0g ….. My local map tiles went from 500KB each to around 1000KB each, in the "stage" subdirectory. I'm gonna try the app tonight to see if we can view them, but just in case some tools need to be updated for this format, or if you can figure out how to take these tiles out for a spin… :D

This is a better question for @DamianXVI I hope he's still around, he's a wonder worker with the maps!
 
  • Like
Reactions: croman
they use the map for driving.
I found this thread because autopilot started doing weird things. Suddenly hitting the brakes (hard!) on a road that's worked perfectly for 12 months.
Lane keeping in areas with bad lines etc.
What's weird is that our AP1 and AP2 car shows different speed on some of the roads.
Different databases?

No, they use maps for localization not for driving, there is a difference. Once, they start to use the maps for actual driving there will be another evolutionary leap in AP IMHO, but not today.
 
  • Like
  • Informative
Reactions: zmarty and croman
It’s not there database to update it’s Tom tom’s, so I bet there is a process to update them, that takes far to long and entirely to much validation so that it’s clearly outdated information before it gets released.
Yup. I've put in update requests, including geotagged photos of the signs. According to the page I found "including a geotagged photo will get quicker response". It was at least 9 months ago. Still wrong on their app as well too.
 
I mean, they are not using maps for driving, they use maps for localization and speeds but not driving, they are using the lanes and markings for driving through the NN, but the maps aren’t saying hey your on a 3 lane highway in the center lane with a 15 degree curve coming in 200 ft etc.

Yes agree, but they aren’t using it yet, just localization stuff which is how they know to slow for curves etc.

Three months maybe, definitely six

No, they use maps for localization not for driving, there is a difference. Once, they start to use the maps for actual driving there will be another evolutionary leap in AP IMHO, but not today.

Wow. Fill me in, what do you mean Tesla is using the maps for localization?

Do you mean Tesla is using the GPS to localize and then using that to extract info from the map?

Or are you implying, Tesla has localization assets in the HD map, that AP2+ cameras recognize to locate themselves?



Also, I am out of date. Are you saying Tesla uses map data to slow down for sharp curves? I did not know this.


Also, I completely agree there will be a major jump in AP performance once the maps are being used.


Amazing, sir. So the branches are called dev, live, stage.

If I look at stage, there's a veeeeery curious map tile name:
na20180212_localizer_optimal: 2018-02-12 15:12



It's newer than those December-ish live tiles we saw, and "localizer_optimal" sounds interesting too :D


EDIT: Paging @DamianXVI , @verygreen , @BigD0g ….. My local map tiles went from 500KB each to around 1000KB each, in the "stage" subdirectory. I'm gonna try the app tonight to see if we can view them, but just in case some tools need to be updated for this format, or if you can figure out how to take these tiles out for a spin… :D

Wait? where do you see the branch names?

And where do you see that localizer_optimal? where do you see that?

that is very intriguing.
 
Wait? where do you see the branch names?

And where do you see that localizer_optimal? where do you see that?

that is very intriguing.

As examples:
curl -k https://daws.tesla.services/iiNG2d6xFR1USl/live/your_geohash.tile | strings | tail -n 1

% Total % Received % Xferd Average Speed Time Time Time Current

Dload Upload Total Spent Left Speed

100 925k 100 925k 0 0 6805k 0 --:--:-- --:--:-- --:--:-- 6854k

na20170502: 2017-06-30 10:05

curl -k https://daws.tesla.services/iiNG2d6xFR1USl/stage/your_geohash.tile | strings | tail -n 1

% Total % Received % Xferd Average Speed Time Time Time Current

Dload Upload Total Spent Left Speed

100 1022k 100 1022k 0 0 4475k 0 --:--:-- --:--:-- --:--:-- 4482k

na20180212_localizer_optimal: 2018-02-12 15:12
 
  • Helpful
Reactions: lunitiks
As examples:
curl -k https://daws.tesla.services/iiNG2d6xFR1USl/live/your_geohash.tile | strings | tail -n 1

% Total % Received % Xferd Average Speed Time Time Time Current

Dload Upload Total Spent Left Speed

100 925k 100 925k 0 0 6805k 0 --:--:-- --:--:-- --:--:-- 6854k

na20170502: 2017-06-30 10:05

curl -k https://daws.tesla.services/iiNG2d6xFR1USl/stage/your_geohash.tile | strings | tail -n 1

% Total % Received % Xferd Average Speed Time Time Time Current

Dload Upload Total Spent Left Speed

100 1022k 100 1022k 0 0 4475k 0 --:--:-- --:--:-- --:--:-- 4482k

na20180212_localizer_optimal: 2018-02-12 15:12


Thank you!

Hmm, only speculation here.

But my guess is that the localizer_optimal, is flagging that tile, or the data / assets in that tile. As high reliability / or optimal... which would tell AP that it can place high confidence or higher weights in the data in that tile. .

or high confidence atleast in regards to localization. since it says localizar_opimtimal
 
Thank you!

Hmm, only speculation here.

But my guess is that the localizer_optimal, is flagging that tile, or the data / assets in that tile. As high reliability / or optimal... which would tell AP that it can place high confidence or higher weights in the data in that tile. .

or high confidence atleast in regards to localization. since it says localizar_opimtimal

I think it's the name of the "branch" that the maps came out of. I don't know whether it means that it's used for localization or some fancy localizer was what was used to generate the map tiles themselves….

At any rate, if you dig for posts by DamianXVI, you'll find a .NET app for opening these tile files. Using that, you can explore all the fields in these tiles as well as the road segments (splines). It's quite fascinating to see.

In general, I'm not seeing a dramatic change in the Feb tiles compared to the live tiles. The "trust" field is bumped up for just about every major road, and the confidence percentages are also dramatically higher.
 
I think it's the name of the "branch" that the maps came out of. I don't know whether it means that it's used for localization or some fancy localizer was what was used to generate the map tiles themselves….

At any rate, if you dig for posts by DamianXVI, you'll find a .NET app for opening these tile files. Using that, you can explore all the fields in these tiles as well as the road segments (splines). It's quite fascinating to see.

In general, I'm not seeing a dramatic change in the Feb tiles compared to the live tiles. The "trust" field is bumped up for just about every major road, and the confidence percentages are also dramatically higher.

Ah, yea I recall that app people were using.

But haven't downloaded it yet. but I did not know there is a "trust" field. thanks for sharing.

is there a "trust" field and a "confidence percentages" filed.. two separate field?
 
Ah, yea I recall that app people were using.

But haven't downloaded it yet. but I did not know there is a "trust" field. thanks for sharing.

is there a "trust" field and a "confidence percentages" filed.. two separate field?

Yeah, there is one field with what looks like a confidence percentage, and another field with just numbers (0/1/2) separate from the restricted road vs non-restricted road (e.g. the 5mph over PSL limit) field. This one was only at 0 and 1 in the current live maps, but even as early as last December, we started seeing the value 2 mark major roads and highways especially around California. The theory is that this somehow either ties into level 3 driving or a reduced nag interval.
 
Wait? where do you see the branch names?
And where do you see that localizer_optimal? where do you see that?
aYrbErj.jpg
 
  • Helpful
Reactions: zmarty and lunitiks
I'm sure they know what IPs should be coming from the LTE/3G in the car. If it was me, I'd feed some super old data to any IPs that weren't coming from the car's connection.
Well you can be tethering over wifi while driving. ADAS tile fetches do go over there when you do so.

But arguably none of us are fetching in the same pattern as the car, or using the right user agent. It should be trivially obvious who's simply scraping it.


They really don't seem to have a reason why they wouldn't lock everything down behind the car's SSL cert, like they do with the dev environment and now the two staging ones.
 
  • Like
Reactions: croman