-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Pentium2004
Former Vendor
Silicon Desert
Well-Known Member
Uh, there is no single download that big. The download must be getting interrupted and restarting.
yangster
Member
That isn't correct, they said that EAP would use two forward facing cameras and the two rear facing cameras.
I'm curious about this as well--hoping to get it while I'm at work. Here's a prior thread on this: Firmware updates only over WiFi?
It has a very specific meaning in engineering: Edge cases are the boundaries of your capabilities, and corner cases are the cases where you are hitting multiple edges simultaneously. But is often used pretty loosely, and I was using it especially loosely in this context to mean that an autonomous system (even L3) needs to be able to handle a vast array of very rare conditions.
For some examples, it needs to recognize pedestrians, which is neither a corner case nor an edge case in itself. But it has to recognize very small pedestrians like toddlers (edge case). And it also needs to recognize pedestrians lying down in the middle of the road (edge case). It also needs to recognize a pedestrian in a halloween costume dressed up as a pumpkin (edge case? corner case? or just weird). And it needs to recognize a toddler lying in the middle of the road dressed up as a fire hydrant (definitely a corner case).
diplomat33
Average guy who loves autonomous vehicles
Well, technically, it would be L2. But as we've been discussing, the levels are not necessarily descriptive of the system's capabilities. You can have a system like Audi's driver assist which is L3 but very limited in when it can operate or you can have Tesla's EAP, especially V9, which is very robust in many situations but is still technically L2. If I am understanding the levels correctly, you could theoretically have a system that is capable of actual full self-driving but is still L2 if it asks the driver to supervise the system while it is self-driving.
i’ve read this example so many times here that i’ve come to a conclusion, that dressing up as a fire hydrant is very popular in the U.S.
That's a valid argument when the computer in your dash in equivalent to the computer in your head; it's not.
To some degree you can make up with a shortfall in processing with additional & redundant sensing.
Did you see the dude on Reddit testing "obstacle aware acceleration" and how you can floor it 2 feet away from another car and it won't crash?
I wouldn't past testing anything past Americans. And let's not even start with that one Summon video where the guy used his small child as a test object for Summon object detection.
You are understanding correctly, and in fact I suspect this is Tesla's fallback plan for their "FSD" option -- to release it as an L2 "self-driving" system. It will drive itself most of the time but you must always be attentive and ready to take over as necessary. It might even nag you to keep your hands on the wheel, though this would really be pushing it I think. (No hands on wheel in the demo video.) But there will be no such thing as operating without a driver behind the wheel remaining responsible (except maybe for some kind of limited driverless self-park and summon features, and honestly I don't know how they're going to weasel through that...)
There may be some limited L3 on the highway at low speed, similar to Audi's system. This means that not only is the Tesla going at low speed, but everything around it must be moving slowly as well. If it detects traffic moving fast in another lane not separated from you by a divider it will fall back to L2 (and require some kind of confirmation from the driver that they're alert again). This is not what most people had in mind when they bought EAP/FSD...
Similar threads
