Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register

Tesla haven't recognized a train

This site may earn commission on affiliate links.
So, I was driving toward train crossing, was expecting tesla to recognize it, but nothing was shown on the screen, what is even worse the crossing red lighted were not detected.
Have you had any experience with train crossings, was your car able to detect them?

FE53F81E-0DC9-4680-8590-FB879BAECEB8.jpeg
 
There's nothing to fix- the current public version of the software isn't intended to recognize or react to train crossings like this.

Correct. And to be clear - just that it doesn’t RENDER the crossing doesn’t mean it doesn’t RECOGNIZE it, either.
Further, I don’t expect any material updates to the current Autopilot software. There’s no point. It’s limited - we know it’s limited - Tesla knows it’s limited - and there’s something better on the horizon.

I anticipate any improvements to Autopilot to come from leveraging the 4D FSD backend once widely released.

Until then - what we have is what we have ... if I were Tesla, I wouldn’t waste one engineer on trying to improve it.
 
Correct. And to be clear - just that it doesn’t RENDER the crossing doesn’t mean it doesn’t RECOGNIZE it, either.
Further, I don’t expect any material updates to the current Autopilot software. There’s no point. It’s limited - we know it’s limited - Tesla knows it’s limited - and there’s something better on the horizon.

I anticipate any improvements to Autopilot to come from leveraging the 4D FSD backend once widely released.

Until then - what we have is what we have ... if I were Tesla, I wouldn’t waste one engineer on trying to improve it.

Plenty of Model 3 out there with the 2.5 computer who will not be upgraded because they aren't buying FSD.

Still needs to be maintained, and train rendering may be added.
 
Plenty of Model 3 out there with the 2.5 computer who will not be upgraded because they aren't buying FSD.

Still needs to be maintained, and train rendering may be added.

As long as the software sees the train, whether the screen renders it or not is not a safety issue.
 
Why?

2.5 Autosteer isn't intended to be used in places with train crossings.

Because people who have EAP and who refuse to buy FSD still believe they are entitled to something other than what EAP currently delivers, which we know they arent.. but many still think that way.

(note, I am agreeing with you)
 
Because people who have EAP and who refuse to buy FSD still believe they are entitled to something other than what EAP currently delivers, which we know they arent.. but many still think that way.

(note, I am agreeing with you)

Correct - and what I said above is " I don’t expect any material updates to the current Autopilot software." That does not include major bug fixes... but I wouldn't expect to see any new features or anything radical for the older software stack.
 
Correct - and what I said above is " I don’t expect any material updates to the current Autopilot software." That does not include major bug fixes... but I wouldn't expect to see any new features or anything radical for the older software stack.

As usual, we agree. The only "splitting hairs" portion of this is exactly what is "bug fixes"? I agree that EAP is complete as a feature set.
that feature set (as @Knightshade points out) will not include navigate on city streets, which is where train crossings would be found.

People will continue to "use it where they are not supposed to" and continue to complain about things not working that are not supported by their configuration, because thats what people do. Maybe tesla provides enhanced visualizations to HW2.5 cars over and above what they have now, maybe they dont.

As @MentalNomad points out, the software will need to be "maintained" but maintenance doesnt include adding features, just ensuring that the current ones work with whatever software version is being pushed out, or fixing bugs when it doesnt.

With all this being said, we have no idea what feature set this OP actually has. Autopilot? Enhanced Autopilot? They dont say. I suspect its autopilot, not enhanced autopilot, and not FSD.


So, if its autopilot, there should be zero expectation that "autopilot" should be used anywhere there would be a train crossing, but... here we are.
 
I'm sure FSD will never detect or render a light airplane landing on the freeway ahead of you. But, I guess there may be drivers that can't 'detect' a railroad crossing.
I am one of those EAP folks (wife has FSD and rarely uses it on her commutes - 35k a year), and I rarely use it, especially on my 3-4 mi daily commute - absolutely no reason for the upgrade.
But, recently did a 5000 + mile tour of the Northwest, and still rarely hit autopilot. I found that when I did, it caused a little more stress (in unusual areas), but definitely made the car drive more comfortably for the passenger.
Bottom line - I think it has come a long way and still has many 'fixes, upgrades to go - and will never replace driver for every situation. But, it will decrease accidents - saved me a couple of times - and I applaud the man for taking on this phenominal task.
 
But I'd day nothing to fix here - this is the FSD visualization preview, which is due to be completely replaced anyway.
Unless there is somehow an (unrealistic) expectation that this will be provided to cars without FSD.
Without FSD it will never stop for any red light so why does the display matter, even less of a concern that it doesn't show a train.
Sounds more like an unrealistic expectation
 
  • Like
Reactions: Ol'Blue
The list of things it currently (NOT including Beta FSD) recognize or at least render is SHORT. The list of things it doesn't recognize/render is LONG.
Please document.

Considering how many things it avoids (all of them so far) and how many things it has run into (none), I say my car does a pretty damned good job.