Auto steer Not ready for prime time

[apsen]

The functionality is still (or seems to be) light years beyond what other cars have right now.

Wow... Really?! Which ones did you try?

 

[KirkNason]

I am on a 30 day trial of the auto steer so I took the opportunity to try it out on a 700+ mile round trip from Phoenix to San Diego. The route is nearly all interstate 75mph in Az and 70mph in Ca. Observations:
1. The "put pressure on the wheel" occurs about 3x/mile a real pia
2. When you go by a ramp entering the freeway and you lose the right lane line the car will veer to the right trying to center itself in the temporary extra wide lane. Unsettling, especially for the passenger who is seeing the car move right toward the shoulder.
3. Lane change was the only real excitement of the trip. Sometimes the car would go into a wobble moving left then right over and over not knowing if it was safe to change lanes. Only way to stop it was grab wheel and disengage. The color of the lane marker seemed to set it off. One time this happened with freshly striped black asphalt. This alone was a deal breaker.
4. When it did successfully change lanes, I found the rate of change too fast. Not good for tire wear and uncomfortable for passengers.
5. Self steering was very good when not changing lanes as long as there were visible stripes on each side of the road.

Don't know if Tesla engineers monitor this site or data automatically uploads to Tesla when a failure occurs but these items need to be addressed. It's fun to try out these new features but there comes a point where you have to question using your customer base as beta testers.

On tripe is way not enough to understand the benefits. I have gone thousands of miles on numbers cross Western US strips and would never ever consider not having this amazing capability. AP is constantly learning and if you had tried in with AP1 you would see how it's constantly improving. Understanding left freeway exits, slowing down for curves, understanding truck shadows, and other things.
Did you submit a bug report when the anomaly occurred? If not, get on board and contribute to making the experience better.
Press the mic button say "Bug Report" and state your concern.

 

[apsen]

4) It's a bit jerky. It hit the brakes first and then dives into the next lane.\

It will either hit the brakes or accelerator depending on how fast that lane moves... Or just changes lane if the speed adjustment is not needed.

 

[apsen]

...or having to disengage when going through 15 MPH school zones because the minimum is 18 MPH.

That's probably why the limit is 18 mph so you don't try to do that.

 

[apsen]

What's interesting and kind of almost scientifically intriguing for me is that there seems to be almost no more divisive subject on the form then reports about autopilot behavior. What does it mean that a significant fraction owners report exactly this kind stuff and another large and perhaps smaller but still significant fraction vigorously disagree? I have to wonder if we're getting different functioning across different builds of the autopilot system? I have no knowledge of differences in camera systems or central computing systems, (although there is clearly updating of this Autopilot code across versions of the operating system – and that might explain some of these variabilities), but one would think with cars running the same version of the operating system and having what appear to be virtually identical camera systems that you could not get disparate behavior out of the system.

Perhaps that's just not true! I would challenge the folks who are disagreeing with this post to provide chapter and verse from their own experience with the autopilot that contradicts specific points here. I've observed much of this kind of difficulty and it's part of why I don't have high levels of confidence particularly when the autopilot system has to manage exits and other transitions. But again it may be that some owners are simply not getting this problem manifested – and those owners should step forward and clarify that their system is functioning differently. If that is confirmed, then Tesla has a major mystery on their hands.

I think it is mostly personal perception. Whatever reason behind it is. You could find people mentioning the same facts but having opposite impressions.

Some people never ride in the cars as passengers and freak out when it makes perfectly fine maneuver just not exactly to their liking, some fail to comprehend the limitations and freak out when AP encounters those limitations, some just plain trying to use it where they shouldn't, etc...

I think this is something akin to uncanny valley - the system is so good that some fail to see it as a mere tool and expect it to behave more human-like and when they see the difference it scares them much more than less perfect system would...

 

[apsen]

I bought it and I own it and indeed across two separate vehicles with the same system I've observed exactly these problems. So I'm afraid your formula does not explain my results.

That's a bias - some people are more affected by it some less. Then again there are a lot of different biases and maybe you are biased against it

 

[apsen]

My hands were not heavy enough on the wheel so I had to grab the wheel- sometimes a little obnoxiously so. I tried auto pilot here and there and it was the same everytime. So to me, as a woman who may not be as heavy as some of you men, it was not worth it to me.

It is not a matter of how heavy you hand are but more how much resistance they have against AP steering. I noticed that when I hold both hands on the wheel I get nags more often than when I use only one hand. I guess I unconsciously pull the wheel and with one hand there's no opposite rotational moment to compensate.

 

[apsen]

Agree, current auto pilot is more stressful and requires intense concentration to use in many situations. Maybe the new CPU upgrade next year will make it more comfortable than a human driver.

This would happen if you are not yet familiar enough with it. Once you familiarize yourself you would know when you could relax more and when you need to pay more attention. My trip from from NJ to Toronto and back was much less stressful than in an ICE car we used to take. After driving back I usually just go to bed but this time I went doing stuff around the house as I was not as tired as usual.

 

[CELSIE19]

I love to Beta test, heck I would Alpha test. That's what it is, a test. So keep alert and note any perceived issues as you have done. Don't fear giving positive feedback either. I love the auto features, however I seem to be inept at the self parking.

 

[Timothyj999]

You can use the scroll wheels instead “of putting pressure” on the wheel.

Agreed that there are some shortcomings but I use AP all the time and find it to be a fantastic feature. Well worth the investment, for me.

I’ve found that simply resting my hand on the wheel, so the car (apparently) detects a tiny bit of resistance, is enough to completely prevent the hands off warnings.
I am in agreement that there are a few exciting moves at lane splits, but overall I find long trips much less fatiguing.

 

[whatthe2]

I’ve found that simply resting my hand on the wheel, so the car (apparently) detects a tiny bit of resistance, is enough to completely prevent the hands off warnings.
I am in agreement that there are a few exciting moves at lane splits, but overall I find long trips much less fatiguing.

Yes, this is a great solution as well.

 

[Ficheh]

I got a 7 or 14 day trial back in Sept, TACC worked fine. Auto steer never worked, the steering wheel icon was always greyed out. I went on a 400 mile (rt) trip and still it never engaged. I emailed Tesla about it, their reply was that they'd look into it. Still haven't heard back. I'd be willing to pay for the TACC, but at a much reduced price since the rest of the EAP didn't work. I'm kinda disappointed they never followed up. Other than that, the M3 has been the best car ever.

 

[Gregwhiteshark]

I also recently finished a 30 day trial. Here was the dealbreaker for me: While passing a car on the interstate while in autosteer, my Tesla 3 suddenly braked very hard for no apparent reason. I wasn't expecting it and it actually tweaked my neck. I had a sore neck for a few days afterward. I didn't use autosteer much after that.

 

[whatthe2]

I also recently finished a 30 day trial. Here was the dealbreaker for me: While passing a car on the interstate while in autosteer, my Tesla 3 suddenly braked very hard for no apparent reason. I wasn't expecting it and it actually tweaked my neck. I had a sore neck for a few days afterward. I didn't use autosteer much after that.

Sorry to hear. I’ve had instances where it brakes hard for one reason or another, but nothing that caused me to not use it. I still find the pros outweigh the negatives.

 

[spinbackwards]

I drive on I-70, Western Slope to Denver.

I use it, but I'm careful.

It doesn't work when it's snowing. I don't trust it when there's snow on the road. Going around curves it gets too close to the lane lines, makes me think I'm going to crash into the car next to me.

Yeah, it's definitely early version beta.

TACC on the other hand, is reliable.

 

[Woodleslie]

I am on a 30 day trial of the auto steer so I took the opportunity to try it out on a 700+ mile round trip from Phoenix to San Diego. The route is nearly all interstate 75mph in Az and 70mph in Ca. Observations:
1. The "put pressure on the wheel" occurs about 3x/mile a real pia
2. When you go by a ramp entering the freeway and you lose the right lane line the car will veer to the right trying to center itself in the temporary extra wide lane. Unsettling, especially for the passenger who is seeing the car move right toward the shoulder.
3. Lane change was the only real excitement of the trip. Sometimes the car would go into a wobble moving left then right over and over not knowing if it was safe to change lanes. Only way to stop it was grab wheel and disengage. The color of the lane marker seemed to set it off. One time this happened with freshly striped black asphalt. This alone was a deal breaker.
4. When it did successfully change lanes, I found the rate of change too fast. Not good for tire wear and uncomfortable for passengers.
5. Self steering was very good when not changing lanes as long as there were visible stripes on each side of the road.

Don't know if Tesla engineers monitor this site or data automatically uploads to Tesla when a failure occurs but these items need to be addressed. It's fun to try out these new features but there comes a point where you have to question using your customer base as beta testers.

I have to agree with you. Since the update I have felt the reliability to be far worse. Even some of speed limit signs are not reading correctly and they were before the update. I unfortunately will be using it less and when I do certainly being more aware until some of the bugs are worked out.

 

[BG10]

That is your opinion - and yet thousands of us use it every single day and love it.
Nothing wrong with having an opinion

I agree. His opinion seems slanted towards justifying why he did not purchase EAP.

 

[bluedog]

All the experiences enumerated here are what i've noticed since taking delivery of our Model 3 in April. The Version 9 upgrade really was an improvement. I now use Auto Steer all the time on our local roads. When they fan out into confusing multiple lanes, i've learned to simply let the M3 handle it (with my vigilant hands on the wheel) and take the sometimes sudden steering in stride. Auto Lane Change even works on our local four lane divided roads. I've not grown complacent, and if anything i feel MORE engaged in the driving experience than if i was driving. Since i drive the same roads with the "multi-lane confusion" every day, i would love to see the software Grow into predicting these situations. Also, i'd like to see the system look further ahead to recognize what's about to happen.

 

[wcorey]

I think it is mostly personal perception. Whatever reason behind it is. You could find people mentioning the same facts but having opposite impressions.

Some people never ride in the cars as passengers and freak out when it makes perfectly fine maneuver just not exactly to their liking, some fail to comprehend the limitations and freak out when AP encounters those limitations, some just plain trying to use it where they shouldn't, etc...

I think this is something akin to uncanny valley - the system is so good that some fail to see it as a mere tool and expect it to behave more human-like and when they see the difference it scares them much more than less perfect system would...

It's call confirmation bias.

 

[dogphlap]

I have a 2015 Model S with AutoPilot (the original MobilEye system).
This has tried to kill me once by sending me into oncoming traffic through a gap in the median strip.
This has saved me once by apparently leaping sideways when a truck moved over into my lane. The truck realised its error almost immediately and went back to its own lane but I'm sure without AutoPilot there would have been an accident, possibly fatal. AutoSteer reacted incredibly quickly, before I was even aware of the danger.
Anyway I now have version 9.
Prior to v9 I had a tick box that allowed me to disable lane change on signalling. I now do not have that option.
I like to decide how early I signal depending on road and traffic but my car now changes lane almost immediately on using an indicator with no options to change from MadMax mode or whatever it is.
I find this behaviour intolerable so I now have AutoSteer disabled which is a shame because the car does a much better job than me of staying centred in its lane.
I have requested the tick box be reinstated and I did get an acknowledgement but I'm not holding my breath.