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

Petition to unlock full Autopilot in EU

This site may earn commission on affiliate links.
Lnae change assist is implemented in other premium german brands for a while now, and assuming they have to follow the same regulations, seems to be working correctly. At least I Haven't read any complaints online.

My guess is that they will need to step up the implementation, coding and software for the EU market, as a separate entity from the US, and this has not been done yet as they are looking to force regulators to adapt to the new technology rather than waste software development time.

Watching a Fully Charged review of the ID Buzz, one of the VW spokesmen clearly said that the self driving is ready but it can't be implemented and they are waiting for regulation, my suspicion is that the proposed changes are being more pushed by the German automakers than Tesla and that's why it is moving
 
  • Like
Reactions: Durzel and MrBadger
I would say that's a fair assessment.

Like everything in politics there is lobbying going on. When the big hitters (VW, Mercedes, BMW, etc) have something ready to go and feel like they are losing out because of these regulations, that's when I imagine things will actually start moving, not because of some "upstart company" in America.
 
  • Like
Reactions: pow216
The last UNECE 79 committee meeting last month discussed increasing the timeout to 15/20 seconds (depending on which amendment they eventually accepted). They haven't published their results yet though and there was a bunch of other stuff in there including a legal framework for smart summon (private land only, basically, which is fair enough).

This presentation from the meeting is fairly interesting.. you might recognise the car they used.. http://www.unece.org/fileadmin/DAM/trans/doc/2019/wp29grva/GRVA-04-53e.pptx

(If you've not got a pptx viewer I converted it using google - GRVA-04-53e.pptx).
 
  • Like
Reactions: Dario_75D
I've had my car for almost a month now and the more I see of AP, the less impressed I am. Phantom braking, autosteer failing to engage, loss of cameras due to dew, rain or dirt, automatic lane correction activating when not needed, awkward lane changing with AP on a highway, autosteer limited in the radius of the curve it'll take, jerky steering around curves, overly abrupt braking for zero risk events like cars turning across your path, inability to handle parked cars, confusion over some road markings like bus stops, ... The list goes on.

I can't believe that anything but a fraction of these are due to EU / UNECE regulations. It's just a bad implementation. I find it hugely ironic that Tesla is known for its fancy auto driving, but in fact it's inferior to many other manufacturers out there.

Thankfully, I didn't buy the car for its AP and accordingly my expectations of it were low - so it's not a deal breaker for me. But really, before we open AP/FSD to more autonomous driving, Tesla need to up the game significantly.
 
My 3 is in at the MK Service Centre to rectify the faults noted on delivery

I have always disliked ‘Adaptive Cruise Control’ but the system fitted to the Model 3 is totally unfit for purpose and I agree with PDK42 in relation to the noted shortcomings - I have AP, not FSD.

What is interesting is the 2015 Model S loan car they’ve given me has adaptive cruise control that actually seems to work as I would expect it to, almost none of the issues noted by PDK42 are evident and it accelerates as you indicate to change lanes on the motorway (the 3 usually hesitates and more often that not actually slows down), even when there is no other cars anywhere near me, also, the driver initiated lane change using the indicator without cancelling auto steer makes the car much more fluid on the motorway in general.

My commute is a 140 mile round trip daily, almost all on the motorway and the older, perhaps more developed tech on the Model S (it’s a pre facelift without side cameras) does the job a million times better than my 3 does.

I suppose we just have to hope that as AP develops it gets at least as good as it is on the old Model S !
 
With all due respect @pdk42 many of your issues are in scenarios that fall outside AP use parameters. During Heathrow handover, the presentation specifically reiterated that AP is only for use in limited scenarios. The fact that it does work in these scenarios is a positive. Tesla could simply have geo fenced the functionality to the roads it does support and played AP nanny but I think that would have been to the detriment of many. Please feel free to criticise AP’s failings when used as designed but I feel criticism of its failings in scenarios where it is not meant to be used in its present form is unjustified.
 
With all due respect @pdk42 many of your issues are in scenarios that fall outside AP use parameters. During Heathrow handover, the presentation specifically reiterated that AP is only for use in limited scenarios. The fact that it does work in these scenarios is a positive. Tesla could simply have geo fenced the functionality to the roads it does support and played AP nanny but I think that would have been to the detriment of many. Please feel free to criticise AP’s failings when used as designed but I feel criticism of its failings in scenarios where it is not meant to be used in its present form is unjustified.

Well maybe you're right on the smoothness of lane changing, but on all the other points I made, I don't see how FSD could be better than AP:

- Phantom braking on basic cruise control - pretty much any other car on the market today does a perfectly good job of adaptive cruise control. I can't see how this is anything to do with FSD.

- Spurious activation of emergency lane departure. This is a feature that defaults to on all the time. It's reasonable that it should work.

- Jerky auto steering - surely they don't program the steering to be jerky on AP but not FSD? I'd find that petty hard to believe.

- Problems with cameras being obscured. Again, are you saying that the software is knobbling AP but it works with FSD?

And even on the lane change, there are lots of cars now with lane following and I'm pretty sure that most (if not all) of them do a really good job to automatically re-engage when the car is back it it's new lane.

I think defending AP on the Model 3 is defending the indefensible.
 
  • Like
Reactions: drewpost
I've a question for those with FSD on the Model 3.
When you have the upgrade but have not activated NoA for a journey, but are using standard Autopilot, does lane changing still cancel Autopilot?

I borrowed a Model S for a little while, and the lane changing on Autopilot is what I would like on the Model 3.
 
The last UNECE 79 committee meeting last month discussed increasing the timeout to 15/20 seconds (depending on which amendment they eventually accepted). They haven't published their results yet though and there was a bunch of other stuff in there including a legal framework for smart summon (private land only, basically, which is fair enough).

This presentation from the meeting is fairly interesting.. you might recognise the car they used.. http://www.unece.org/fileadmin/DAM/trans/doc/2019/wp29grva/GRVA-04-53e.pptx

(If you've not got a pptx viewer I converted it using google - GRVA-04-53e.pptx).

It'll be nice if they accept the proposed amendments in that .ppt - pages 12 and 13 show that the 5 second auto lane change limit is a killer. The proportion of attempts that are aborted is huge. User experience would be improved massively by taking that restriction away as we all know from using it. Plus the restriction surely decreases rather than improving safety because it creates erratic moves.

Personally I am not convinced there's a reduction in phantom braking from the move to v10 either. I'm still getting quite a lot of it.

I also think this weaving in the lane thing is poor. And I think it's new. I suspect it's an unintended consequence of the new adjusting your line to pass a lorry thing.

Deal with these three negatives together and the driver/passenger experience would be improved hugely. And surely none of these is hardware-limited.

My other frustration is its bad habit of slowing down unnecessarily for a lane change. It's bad driving and uncomfortable for the passenger. I agree that it didn't seem like this on pre-face-lift Model Ss.
 
Last edited:
I've a question for those with FSD on the Model 3.
When you have the upgrade but have not activated NoA for a journey, but are using standard Autopilot, does lane changing still cancel Autopilot?

I borrowed a Model S for a little while, and the lane changing on Autopilot is what I would like on the Model 3.
No. Lane change keeps you in AP (strictly, it’s “Auto Steer”, activated by the double-tap ontop of the TACC single tap.

NOA is a per-drive option available when Auto Steer is active
 
My gripe is not the criticisms when AP should work, but the criticisms when it is being used outside its design parameters or when limited AP functionality is criticised when that functionality was/is part of the EAP/FSD package. If its not being used on Motorway or segregated major carriageway then its a case of Caveat Emptor.

I've had my car for almost a month now and the more I see of AP, the less impressed I am. Phantom braking, autosteer failing to engage [on a motorway?], loss of cameras due to dew, rain or dirt, automatic lane correction activating when not needed, awkward lane changing with AP on a highway, autosteer limited in the radius of the curve it'll take [on a motorway?], jerky steering around curves, overly abrupt braking for zero risk events like cars turning across your path, inability to handle parked cars, confusion over some road markings like bus stops, ... The list goes on.

I'm far from defending AP shortcomings when criticism is justified, but this is a thread about AP being restricted by regs and many of the restrictions relate to functionality in EAP and/or FSD where more cameras are/will be utilised than standard AP so yes, standard AP is knobbled if you like, or maybe another way of thinking about it is that FSD functions are potentially more resilient - I have no official source for this info about camera usage but there are various discussions on it elsewhere.

Lots of improvements needed for AP I agree, but some will only come to FSD.
 
I wholeheartedly disagree. I use NoA every day driving from home to work. It's a joy. Most of my journey is on the motorway. It handles lane changes and exits really well and autosteer handles queuing/slow traffic perfectly.

I agree AutoSteer is great. The issue is with the added value of NoA over and above AutoSteer. Apart from lane change, there doesn't seem to be any.
 
As I was fast-tracked and never got to see the presentation, can you remember what those scenarios are? Thanks
Much the same as the text you must have studiously read and agreed to when you activated it in your car :)

autopilot_disclaimer.png
 
Some of the common failings do happen in its design parameters. I've had autosteer unavailable for periods on a motorway (eg apparently for front facing radar not being able to see well enough, on a stretch of a few miles of a single right hand lane temporarily on the opposing carriageway through roadworks). Plenty of phantom braking too and the weaving thing. And of course the aborts and the inhibiting of auto lane change for the mildest of weather.

However, I hope some of these will improve dramatically with an update or three.

And I do regard it as a generally decent system and real positive of the car which I use a lot.

There are also cases outside its design parameters where it is really useful, but you shouldn't expect them to be OK. I don't think they're common. One single carriageway A road I use regularly, typically in a stream of c.50mph traffic, it does a really solid job on.
 
  • Funny
Reactions: Strategium