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

Cruise control no longer safe

This site may earn commission on affiliate links.
Lucky you. I don't count single carriageway instances. My two most severe experiences were on the M6 and the M40 and I also have follow at 5 spaces. M6 was perhaps understandable in as much as plenty of traffic but on the M40 case there were no bridges or traffic in the middle lane and fortunately nothing behind.
I travelled M6/M5/M42/M40 today (collecting family member from medical treatment) & had three instances down & one back. The worst was M5 crossing Junction 1 with very rapid deceleration from 70 to 30 even though my car settings are for 'current speed' where a normal car would continue at that speed. One of the others was with a curtain sided lorry but quite gentle.

Driving with foot hovering over the accelerator is quite uncomfortable on long journeys so I turn Autopilot off more than on.
 
But I still get caught out by bridges and I can't figure that one out.

Have you seen verygreen's videos showing radar reflection clouds when passing under bridges? And another where some people walking around well to the left seemed to have a cluster of reflections mimicking their movement but more central in the visualisations?

I'll see if I dig up examples. They are from some time ago, but not much changed in car behaviour in this regard since then.

For anyone that follows Green on Twitter it’s interesting he has experienced this recently also

green on Twitter

That is common for me in UK, often near motorway junctions but definitely NOT only where there have been changes to road layout / speed limit any time in recent years.



Exactly same experience. Regarding overhead signs or large flat side of trucks / buildings, I believe it is likely due to radar reflection.

See this clip at the start point here: (at 0:28)


then watch from a little earlier. The green dots on the far left of image look like they are related to the pedestrians in the centre of the image.

Greentheonly also has posts of passing under bridges and seeing radar detection points cloud up on the road surface which could easily be read as an object in front of the car.
 
  • Informative
Reactions: akenham
Have you seen verygreen's videos showing radar reflection clouds when passing under bridges? And another where some people walking around well to the left seemed to have a cluster of reflections mimicking their movement but more central in the visualisations?

I'll see if I dig up examples. They are from some time ago, but not much changed in car behaviour in this regard since then.
Thanks I'd be interested.. I had one with a tall fir tree on the way to Huntly, still got that saved on the USB. At that time I thought this was its shadow on the road causing the problem.
 
Thanks I'd be interested.. I had one with a tall fir tree on the way to Huntly, still got that saved on the USB. At that time I thought this was its shadow on the road causing the problem.

I can't say it isn't shadows. Enough people mention shadows seeming to be linked to unexpected braking, but presumably in order to have a shadow you need something to cast the shadow.
 
  • Like
Reactions: exlatccatsa
I can't say it isn't shadows. Enough people mention shadows seeming to be linked to unexpected braking, but presumably in order to have a shadow you need something to cast the shadow.

My experience of AP and a couple of bridges on M3 is that the phantom braking events only occur on days where a distinctive shadow is cast. Same bridges, different lighting, different AP/TACC behaviour.
 
  • Like
Reactions: Jason71
My experience of AP and a couple of bridges on M3 is that the phantom braking events only occur on days where a distinctive shadow is cast. Same bridges, different lighting, different AP/TACC behaviour.

Then maybe it is a mix of radar and visual, and a strong shadow in the right place relative to the bridge structure gives the false response? I believe that is likely a different scenario to flapping curtain sides etc, but both examples of the present system obviously failing under certain circumstances.
 
Last edited:
I have the model 3 2020, updating always the software. Very bad experience with the TACC as well.
I find it unreliable and dangerous the way it is at the moment. I have many bad examples but I will mention the one that shocked me mostly: I was driving in the highway about 120km/h and passed under a bridge. It was during the day so there was some shadow under the bridge on the road. The car decelerated massively. I believe had a car being behind me it could have led to a potential problem.
I decided not to use it any more.
Tesla needs to fix that asap.
 
  • Informative
Reactions: Matias
Then maybe it is a mix of radar and visual, and a strong shadow in the right place relative to the bridge structure gives the false response?

I am almost certain that no one type of sensor is being used as a single point of reference.

With FSD city streets, I heard someone (I believe a beta tester) actually say that radar was not being used for city streets FSD (so not the same circumstances as being discussed here and im most certainly not implying that radar is not used for highways) although I cannot recall who/where to find it or yet to notice a second source. Someone did speculate that vision and AI was primarily used with other sources as confirmation, but I cannot recall if other sources meant maps or extended to other sensors such as radar. There were several examples where predictions were wrong such as the car predicting a circular roundabout rather than carpark layout and another where it was visualising a phantom dual carriageway where the other carriageway was actually not there. Hmmmm, phantom objects... where have I seen that before?
 
almost certain that no one type of sensor is being used as a single point of reference.

Likewise. Especially at speed.

However at slow speed (ultrasonics) it could be single sensor. And in the case of certain emergency behaviour, I can imagine maybe basing a decision on a reduced sensor set.

Edit: Not forgetting, of course, those stationery vehicles on dual carriageways (delivery trucks, breakdowns.....) that don't appear to be responded to at all. There are clearly multiple feedback / control envelopes outside of which these systems can't make valid / reliable decisions.
 
Last edited:
I did a long journey recently in this bad weather we have been having and not a single incident with bridges but have had it in sunny weather so I also think shadows or the rapid change in light levels, which is essentially the same thing, are a factor.
I am also with others in thinking that curtain sider HGV's seem to be worse for phantom braking than any other type of vehicle
 
One thing I’ve noticed is that some days are worse than others. I’ve put it down to the weather conditions but I wonder if it could be due to updates....does the calibration get reset after an update? Had a faultless drive today having done quite a few miles after the last update
 
One thing I’ve noticed is that some days are worse than others. I’ve put it down to the weather conditions but I wonder if it could be due to updates....does the calibration get reset after an update? Had a faultless drive today having done quite a few miles after the last update

What conditions to you notice it better/worse than others?

I'm not aware that things get recalibrated after an update, but almost certainly functionality changes from release to release as Tesla change and/or compare operation from one version to another even though no mention in the release notes.
 
  • Like
Reactions: Battpower