EAB fail miserably in winter conditions, Tesla to the rescue ?

[Ande]

There have been a major AEB test in Finland, of course, they skipped involving Tesla cars...

The take from the test is that AEB kind of worked at 20km/h for Audi, Ford, Lexus, Mercedes .. they all failed to account for cold weather/traction.
Worst was Volvo, Volvo needed 10km/h to work.. except for BMW X2 and Hyundai Kona who failed at even 10 km/h.

Audi A1, Hyundai Kona and Jaguar E-Pace did not even warn the driver, just crashed into the test "car"


Now: there will surely be more such tests in the future, and Tesla , with impressive traction control, and sophisticated autopilot, should be able to add needed safety margin to have working AEB.

this small victory is easily up for grabs by Tesla, who can update OTA if needed.

 

[Ande]

you may need google translate to understand Norwegian, but here it is: https://www.dinside.no/motor/nodbremsen-svikter-deg-pa-glatt-fore--kan-gi-krasj/70815886

 

[The Rookie]

Given the Tesla manual tells you they will crash into stationary vehicles when in AP under certain conditions, I don’t think Tesla AEB is currently any better than the opposition!

 

[Tam]

...Now: there will surely be more such tests in the future, and Tesla , with impressive traction control, and sophisticated autopilot, should be able to add needed safety margin to have working AEB.

this small victory is easily up for grabs by Tesla, who can update OTA if needed.

It's speculation.

It's just like saying a well-trained gymnast will win the Olympic medal.

 

[Ande]

No, I do not speculate that it already is much/better I just state the the advanced tracking control could be used to automatically test traction in low temperatures, and thresholds for EAB adjusted by that.

 

[The Rookie]

And the other cars don’t have advanced traction control?

 

[Ande]

And the other cars don’t have advanced traction control?

Those are dumb automatic standalone systems, no central processing of feedback data.
Nor can they be updated over air with such feature

 

[The Rookie]

No they aren’t dumb stand-alone systems, in a word CAN.

It’s good to be a fan, being blinkered isn’t so good. How does OTA help still has to be developed first.

Doesn’t work flawlessly

 

[Tam]

No, I do not speculate that it already is much/better I just state the the advanced tracking control could be used to automatically test traction in low temperatures, and thresholds for EAB adjusted by that.

You might have over-designed Tesla EAB. The manual says it is not designed to apply brakes to a halt. It also says the best it can do is to make the collision softer as opposed to full force:

If you doubt about its warning, it gives an example:

If your car is about to collide at the speed of 90 km/h, it will apply brakes to slow down to 40 km/h THEN it will no longer brake so your car can collide at a reduced speed of 40 km/h instead of 90 km/h.

 

[Ande]

Yes, that is so to reduce consequences of false positives, and at speed, due to radar range, and finally not to promise more than can be expected.

Still, easy room for improvement.

Let engineers solve it, not try to discuss it here.

 

[Tam]

...Let engineers solve it, not try to discuss it here.

Yes. Tesla needs to solve collision problems in order to have an unsupervised FSD.

There have been numerous reports of crashes from very slow crawling speed Summon/AutoPark to high speed highways.

We can argue that these features are not FSD but Tesla has just re-classified the following as FSD:

Navigate on Autopilot
AutoPark
Summon
(among 2 others coming out this year)

There's no promise of when unsupervised FSD will happen.

So, for the purpose of this thread, it is still up in the air on whether Tesla AEB will win the competition of braking to a halt to avoid collisions.