I would hope mass means that mobileye may be able to start selling it to any manufacturers who wants it and not just some small company Zeekr.
-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Elon: FSD Beta tweets
- Thread starter Terminator857
- Start date
diplomat33
Average guy who loves autonomous vehicles
I agree. I am hoping too that Mobileye brings L4 to consumer cars in the US by 2025. If the L4 was available on a good and affordable EV, I would definitely make the switch. I trust Mobileye's FSD more than I trust Tesla's FSD.
Actually, the saying is "no HOLDS barred," as in wrestling holds, but to be honest, I think I like your version better.
I remember reading that China has one of the youngest road networks in the world, especially for highways etc.
I find myself wondering if that would equate to more consistent road signs, conventions etc - maybe that makes it easier for autonomy ?
Don't know enough to say one way or the other.
In the US and certainly in Texas there isn't even standardized road marking across the state, let alone the entire country.
I always remember the traffic circles in New Jersey, where "local custom" applies to right of way
diplomat33
Average guy who loves autonomous vehicles
I remember reading that China has one of the youngest road networks in the world, especially for highways etc.
I find myself wondering if that would equate to more consistent road signs, conventions etc - maybe that makes it easier for autonomy ?
Don't know enough to say one way or the other.
In the US and certainly in Texas there isn't even standardized road marking across the state, let alone the entire country.
I always remember the traffic circles in New Jersey, where "local custom" applies to right of way
Certainly, better road infrastructure can help autonomy. It's one reason why Tesla's FSD Beta works better on clean roads with clear road markings. But in the case of Mobileye, I don't think it should make a big difference since they use HD maps. Road markings and signs are added to the map. So the HD map will help the car regardless of how good the road infrastructure is.
right - it always baffles me why Tesla don't use better maps, especially when so many issues we see in FSD beta are due to bad map data.
I also remember Tesla touting how they had so much better map data with individual lanes etc etc - but my car shows no evidence of that.
More realistic rendering of vehicles in 10.12. Gone is the cyber whistle.
sleepydoc
Well-Known Member
Ok... but how about the ability to stay in a lane?
10.12 release notes
Tesla FSD Beta v10.12 Release Notes (2022.12.3.10) - Includes training data from 250k real-world video clips and new vector vehicle assets
Tesla FSD Beta v10.12 Release Notes (2022.12.3.10) - Includes training data from 250k real-world video clips and new vector vehicle assets
- Upgraded decision making framework for unprotected left turns with better modeling of objects' response to ego's actions by adding more features that shape the go/no-go decision. This increases robustness to noisy measurements while being more sticky to decisions within a safety margin. The framework also leverages median safe regions when necessary to maneuver across large turns and accelerating harder through maneuvers when required to safely exit the intersection.
- Improved creeping for visibility using more accurate lane geometry and higher resolution occlusion detection.
- Reduced instances of attempting uncomfortable turns through better integration with object future predictions during lane selection.
- Upgraded planner to rely less on lanes to enable maneuvering smoothly out of restricted space.
- Increased safety of turns with crossing traffic by improving the architecture of the lanes neural network which greatly boosted recall and geometric accuracy of crossing lanes.
- Improved the recall and geometric accuracy of all lane predictions by adding 180k video clips to the training set.
- Reduced traffic control related false slowdowns through better integration with lane structure and improved behavior with respect to yellow lights.
- Improved the geometric accuracy of road edge and line predictions by adding a mixing/coupling layer with the generalized static obstacle network.
- Improved geometric accuracy and understanding of visibility by retraining the generalized static obstacle network with improved data from the autolabeler and by adding 30k more videos clips.
- Improved recall of motorcycles, reduced velocity error of close-by pedestrians and bicyclists, and reduced heading error of pedestrians by adding new sim and autolabeled data to the training set.
- Improved precision of the "is parked" attribute on vehicles by adding 41k clips to the training set. Solved 48% of failure cases captured by our telemetry of 10.11.
- Improved detection recall of far-away crossing objects by regenerating the dataset with improved versions of the neural networks used in the autolabeler which increased data quality.
- Improved offsetting behavior when maneuvering around cars with open doors.
- Improved angular velocity and lane-centric velocity for non-VRU objects by upgrading it into network predicted tasks.
- Improved comfort when lane changing behind vehicles with harsh deceleration by tighter integration between lead vehicles future motion estimate and planned lane change profile.
- Increased reliance on network-predicted acceleration for all moving objects, previously only longitudinally relevant objects.
- Updated nearby vehicle assets with visualization indicating when a vehicle has a door open.
- Improved system frame rate +1.8 frames per second by removing three legacy neural networks.
Last edited:
I'm hoping this version will improve the completely random and unnecessary lane changes when I'm 50 feet from my turn.
JupiterMan
Active Member
I wonder if falcon wing doors will be recognised and visualised differently?!
It probably will get the speed bump if it is in 12.3.11 (no notes on it yet). Looks like we are getting an up to date version. It is not in the Beta notes because it is NOT in the FSD Beta Stack. The notes shown are only about the Beta Stack. You just probably need to scroll down past Beta to see other improvements like AP's 85MPH.
EDIT: That brings up a question: How fast will the Beta Stack go? Will it go past 60MPH (think that is about as fast as I may have ever gone on Beta Stack)? Know it is not safe to test usually but how high of speed can you scroll the wheel?
Last edited:
I think I've done +/-65mph in a 55mph two-lane divided road (where people generally do 65-70) that wasn't considered a "limited access highway" despite looking like one, so it ran the beta stack as opposed to the highway one. Worked just fine. I will try to see how far it'll let me set it next time I'm there.
Daniel in SD
(supervised)
Chuck thinks Tesla has been testing the infamous turn. I think it was probably a fan though...
My biggest concern has continued to be the car crossing the double-yellow line into oncoming traffic to go around a "parked" car in front of it (even though the car isn't parked - it could be waiting to make a left or just the last in a line of a dozen cars waiting at a light). Looks like this next version is claiming to have solved 48% of these cases?!? So it will try and kill me half the time, I guess.
I can tell you how to solve this problem 100% of the time: DON'T CROSS A DOUBLE YELLOW LINE!
I can tell you how to solve this problem 100% of the time: DON'T CROSS A DOUBLE YELLOW LINE!
I just posted about the same response to this in another thread.
FSD Beta 10.12
These all sound very encouraging, but as always we'll have to see how they work out for various people and locations. Our favorite UPL tester Chuck Cook has been out of the FSD beta action with his car in the shop after the very unfortunate accident he had (tragic situation but nothing to do...
teslamotorsclub.com
Similar threads
