you did not what?!
Can the cameras installed on the car, or any car on the market for that matter, do that? No, I didn't think so.
And that's why everybody else uses a combination of different sensors so that if one is not working another can pick up the slack.
Your own response to glare is the solution..drop the vizor, if that's no good then shade a hand over your eyes and drop gaze, drive slower. Until Tesla have redundant cameras, shade, polarized , tube extrusions etc and lens cleaning, heating systems in place for all cameras then they are just using drivers as cheap testers and fodder for their learning process.
i do 20k+ a year. most of it on auytopilot.I've never seen any of those messages in my MY, and I've logged many thousands of miles with Autopilot on. At nighttime, facing straight into a sunset, in heavy downpouring rain, in snow, on both highways and back country roads (though always with painted lines on both sides of the road).
I'm beginning to wonder if various Tesla's somehow perform very differently from one another...
Are you planning to drive on the sun? If not then I don't see this as a crushing blow. However hopefully we can agree that if cameras are capable of taking photos of the sun they're probably also capable of dealing with having it in frame.
Which makes the vehicle more complex and failure prone, increases power demands for both running the sensors and digesting their output and obviously increases the purchase price of the vehicle. It also introduces the need to reconcile different outputs from those sensors (though I think Musk may have been overstating how big an issue that is). Everything has tradeoffs and the fundamental principle that humans manage in all the conditions we're capable of using vision only. There is no reason why an automated system can't do it either.
No. this just shows the fcuking problem when camera is blocked by the sun for 30% of the year
that is pure BS. ONLY for tesla this is somehow an issue. And not for anybody else. somehow "not a software companies like MB, BMW, VW, KIA and others have no issue mixing multiple sensors for few decades. only smart tesla has issues there.
Theory fine but radar would have to have much, much higher resolution, multiple radars and capable of dealing with things radar are not good for - I doubt the echo off a sheep works too well. In conjunction though is different..
Country roads with white lines each side? That would be a major road in Wales, let along when you get to the impossible single track stuff with overtaking bays (aka mud patches and gate entrances). It's not helped by headlights dipping at every sharp bend with chevrons reflecting back at the car.I've never seen any of those messages in my MY, and I've logged many thousands of miles with Autopilot on. At nighttime, facing straight into a sunset, in heavy downpouring rain, in snow, on both highways and back country roads (though always with painted lines on both sides of the road).
I'm beginning to wonder if various Tesla's somehow perform very differently from one another...
Buddy, if nobody ever reinvented the wheel it'd still be made of stone. Sometimes incremental improvements are not enough and someone has to go back to the beginning to find the next step forward.
But there are significantly more cars sold in the USA than Canada, right? A system trained on vast quantities of data, would learn the most from the most amount of data available? More US data than Canadian would mean it learns more about US roads? The weather is also significantly different in Canada than the UK, particularly the type of snow and dryness of it etc. I have noticed that the Now You Know guys seem to have more issues with rain with their FSD around Boston, since their rain is different to other parts of the USA. I'd guess the UK would have similar issues. Which is my point. The system is going to react differently in different countries with different weather, sun positions etc. So we agree there, I guess.
