Een paar snippers uit een uitgebreide discussie die ik had
I think the default position, until the REM system shows it's capabilities, is for initial maps (based on Lidar, or other sources) to come from a "traditional" map supplier, with REM providing change detection.
---
I think, especially for GM, which has long worked on this stuff, they want to understand this space before they commit, and that may not be possible. I think the Germans are a bit more in the mood for giving it away if it looks like it lets them move faster, but are still cautious.
---
I am a little surprised the VW deal is taking so long. When Shashua was here in xxxxxx xxx he said they were a week or two from the mapping deal, and that the sticking points are around data sharing. Still sounds like a week or two. This must still be a big deal- under what terms will VW share data with GM, etc. I am almost positive that this is still unknown with VW, let alone the other OEMs which are farther behind. OEMs may want to be independent, have a small consortium.
They know they are sharing, just not sure they want to. Mostly they know the data might be worth a lot of money and they don't know if they are getting a good deal or being taken for a ride. Nobody does at this point. They also don't want to help the competition without compensation- either appropriate amount of data back or money, but the data-money conversion rate is not clear.
----
I think the default position, until the REM system shows it's capabilities, is for initial maps (based on Lidar, or other sources) to come from a "traditional" map supplier, with REM providing change detection. At least for the HERE owners, but I expect also for GM and others which are sourcing maps from other sources, but don't have the update capabilities.
The issue being debated is do they get the raw data or a processed Roadbook (and what exactly is in a Roadbook), or some other data set based on MI cameras? I think this will depend very much on the business relationships, and probably some elbow twisting by the OEMs (certainly the HERE owners will push for cooperation between MI and HERE), depending on how they want things to go.
In general, I think many of them are skeptical of HERE's ability to deliver, but Mobileye is unknown in this space. I guess the leaders want cooperation for faster results, followers probably want independent work to have options.
----
GM has put a bunch of money into GeoDigital and they are building maps for GM, at least of the freeway network (and I'm hearing way over budget). Ford is developing Civil Maps, but I have not heard that they are building maps yet.
Geodigital uses traditional collection techniques- I doubt they can maintain the maps (realtime), but that is where all of Onstar's work on probe data will help them out. I don't know about their map engine. The map is becoming important and all car companies would like to control the component. I doubt they can, but it is to early to tell (and they don't have the option to buy right now- doesn't exist).
I think GM really wants to make use of the data and capitalize on all the effort that has gone into Onstar- but I don't think they will be able to. They do underestimate the complexity of the complete map, but , even if they can have a unique layer, or do change detection and avoid certian problems, it will be worth it.
I think, especially for GM, which has long worked on this stuff, they want to understand this space before they commit, and that may not be possible. I think the Germans are a bit more in the mood for giving it away if it looks like it lets them move faster, but are still cautious.