Looks like the rollout has stalled a bit. I hope this is the quiet before the (rollout) storm.
-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Helmuth
Member
I’m HW 2.5 in Florida.
The interesting thing is if the radar shows reduced visibility the cameras are (most of the time) still able to see and show the lane markings on the IC! Might be that cameras are the best choice for tough conditions? Better than radar and lidar? We’ll see how the developers are coming up with for those situations (like heavy rain)...
Very true the vision still tracks the lane lines but it might not recognize cars and that's where radar is essential. You need to fuse sensors to really be safe. I think it should knock the system down to L2 but not disengage it.
Helmuth
Member
Totally agree about the radar to be necessary! But it shows that even radar has issues with heavy rain! How would lidar perform in those situations?
Tesla stated that AP2 was capable of FSD when it was released, and said it would use 8 cameras.
SpudLime
Active Member
What brings you to the conclusion that it’s not public release? If you are under the EAProgram, you can’t be connected to teslifi and /or other 3rd party apps.
Good point.
Wayback Machine is not giving me old results for some reason. Wasn't the wording that the vehicle had the hardware needed for FSD?
Elon said that if the computer couldn't do FSD it would be upgraded.
Tesla CEO Elon Musk says ‘almost all new cars will be self-driving within 10 years’
boonedocks
MS LR Blk/Blk 19”
HW2.0 March 2017 S100D delivery (maybe it's because I pre-paid for FSD Bwha ha ha ha)
Tesla stated so many things that have turned out to be completely bogus, not just about AP2 but throughout their history. In 2016 when they started marketing AP2 they had no idea how they were going to implement any of this this; they had nothing even at the alpha stage of development.
At this point I think they have a pretty good idea what EAP will require, except probably for Smart Summon which has not seen any visible progress. I think they're still largely scratching their heads about how FSD is going to work, just like the rest of us.
There are efforts to apply various mathematical models to screen out the noise for lidar but radar is still the most useful for highways and cheapest sensor. That won't change and Tesla needs radar. They needed to better design it to make it more robust. I wonder if the Conti 2.5 one is more robust?
Bumper
Member
I appreciate the solidarity, but I don’t think enough birds were harmed in the making of this update. I’m still not v9.
Depends on the type.
I think Tesla made the right decision in not going with Lidar initially because at the time there really wasn't an good solution in terms of all-weather functionality, and cost.
But, I think things are quickly changing in the Lidar world.
This one supposedly has no issues with rain.
Two Apple veterans built a new lidar sensor—here’s how it works
The biggest issue I see with the current sensors is they simply can't do what's promised. That's not the fault of the sensors, but of Elon's ego.
Even with V9 I would feel a lot more comfortable with rear radar in addition to the current cameras.
d21mike
Active Member
Curious: If they get booted does that mean we should see cars going from 39.4 back to 36.2? Is that possible? Also, if they wise up and stop logging on TeslaFi will we see the number drop because the cars are not longer "active". Not sure what TeslaFi does with stats on cars that are not longer active logging on their system. Or maybe they are still active because they still have an account but not logging. Again, just curious. I am a member of TeslaFi.
heavy rain? It was barely drizzling when I lost radar.
edit: I even pulled over and restarted the car / cleaned off the area of the front bumper that covers the radar. I did not get radar back until the drizzle stopped
diplomat33
Average guy who loves autonomous vehicles
Agreed. But I do think that as LIDAR improves and the cost continues to come down that Tesla will eventually integrate LIDAR into their next cars in a few years. LIDAR in conjunction with other sensors can provide some good benefits to a FSD system. And it's inevitable because technology is always improving. Just as we got HW2, we will eventually get a HW3 and HW4 at some point. Plus, while the current HW2 is most likely good enough for a robust L2 autonomous system and maybe even good enough for a "human supervised" FSD system, if you want a truly L4 autonomous system, you probably will want a hardware system that uses an "all of the above" approach with multiple cameras, ultrasonics, front and rear radar and 360 degree LIDAR for maximum coverage and redundancy. Heck, even if Tesla does pull off a L4 autonomous system with the current hardware combined with the new FSD chip, it would still make sense to upgrade to a better hardware suite in order to get an even better L4 autonomous system because ultimately if you are going to put thousands of L4 autonomous cars on the road, you don't want "just good enough", you want "definitely good enough". So, what I am saying is that the current hardware is good enough for now but it won't be good enough forever.
d21mike
Active Member
Right now on TeslaFi we have 0 new cars on 39.4 and 48 new cars on 36.2. That would normally reverse once a new version goes out to other cars. So, I believe no one outside of the EAProgram and Rooted cars have it.
Rain is supposed to be one of the strengths of radar compared to both cameras and lidar. How heavy is this rain?
This sounds like something else, like maybe water getting through some seal into the sensor or electronics. I have definitely used the radar in heavy rain with no problems; I think you should talk to service about this if you haven't already.
Similar threads
