Auto steer Not ready for prime time

[IROAR]

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.

 

[Mo City]

Agree with you completely. My 30-day trial was a huge disappointment. Very glad I did not purchase EAP.

Maybe in a year or two.

 

[mswlogo]

It gets real exciting when the lane widens and then suddenly splits and the car is centered on the split.

I noticed a lot of exits now have a dashed line which avoids the “wide lane”.

 

[Twiglett]

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

 

[Krugerrand]

 

[whatthe2]

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.

 

[Runnergirl]

Auto steer was a lifesaving feature for me on a recent emergency 8 hour drive at night. I had zero intentions of keeping EAP after the trial. I was just going to play around for the 30-days, and be done. The emergency drive was a blessing in disguise. I completely disagree with the original post. While I agree future updates will only improve the existing software, the auto steer as it is, was phenomenal. It actually trained me how to use it better. The first hour I hated it. I didn't trust it and also had to grab the wheel and it would disengage. There's a learning curve. I discovered I drive too far left in a lane in general... so it pulls me back to the middle. Hour 4-8 of the drive; I decided EAP is a must and you learn little tricks that make it more like natural driving. I won't even use it on my daily commute to work because I have a long rural route, but knowing it's there for the next road trip is reassuring. I think its worth every penny and adds a safety component that is priceless.

 

[apsen]

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.

Guess it is too early for you to be using it. To each his own. I'm ok with trying it and it does state that it is still in beta...

As to your comments:

1. Not that often to me - I'm not tall but stocky and it seems like the weight of my hands on the steering wheel takes care of it mostly. Also it is necessary evil as a lot of folks would not pay attention (defeat devices?) so we all have to suffer for it...
2. Agree but it seems to be doing it less the more I drive.
3. Haven't experienced it - it may try and go back but I don't remember it doing attempting it again and again. It does seem too skittish and will abort the attempt sometimes - too often if the traffic is heavy enough. I have the collision avoidance (or whatever it is called in the settings) set to 'early' so maybe I need to try more aggressive one and see if that makes any difference.
4. For me it was very smooth. It did it faster than I usually do but slower than needed in heavy traffic (causing #3, I guess... I also take forever to change the lanes given the chance). Ideally it would be better if the speed of lane change would adapt to road conditions.
5. I read about it needing to be able to see the lines before I tried AP so I was actually surprised how well it did it. Some lines that were worn to the point where you almost need to guess at their existence were enough for AP. It does seem to be doing worse if instead of lines on the right it is limited by the stones like if there should be sidewalk but there's no actual sidewalk (sorry, not sure what it is called...). There's at least one place around here where it is getting too close to that side for my comfort that I always take over.
6. When Navigate on Autopilot takes exit it overshoots and kind of bouncing off the line denoting the outer edge of the exit lane. It never crosses that line but gets close enough that rumbles on the reflectors set inside that line. Seems to be worse at higher speed - when I had slower moving car ahead of me forcing AP to slow down it did really well.
7. Was mentioned elsewhere on this forum and I have experienced it once or twice: phantom breaking.
8. Stop and go traffic: seems to take too long to accelerate sometimes if the car ahead of you is able to accelerate fast. Feels like I'm in a low powered ICE car in those cases

 

[lolder]

The lane change has to be quick for use in heavier traffic. It is smoother now than originally. I went to use it a few days ago to pass and there was a car behind me also. I looked, hit the stalk, the car started to change and then abruptly stopped the change. The car in back of me had pulled out to pass me in less than a second. I continued my pass and then the rear car passed me doing about 90 and wove in and out of traffic as it disappeared ahead. The wandering into the center of on ramps is milder than it once was and will probably improve more. Even though it is occasionally a little edgy, remember it is looking all around the car all the time.
Putting "pressure" is not the desired action. It wants torque or turning force. The weight of your hand on either side of the wheel is enough. I have learned to hold the wheel lightly enough that it does not require conscious action and hardly ever get a nag.

 

[Big Earl]

Autopilot is always improving. We use it regularly on long road trips and while it has some annoyances, quirks and occasional freak-outs, its performance overall has been outstanding. You, as the driver, have to know when and where this beta feature works best. As you get used to it, you'll learn what it can do well and what it has trouble with.

 

[PhaseWhite]

I think the OP has pointed out legitimate issues that the current version of AP has... However as someone who's been driving on AP 2.0 for over a year and a half, this is by far the most mature version and it's way better than anything else you can get in a consumer vehicle.

I do really wish that Tesla would find a solution to the issue with the car trying to center on merge lanes where the dotted lines go away as you are driving in the right lane. This is by far my biggest issue when using Autopilot on a 2 lane highway. I either need to be in the fast lane, or in the slow lane and tolerate the car swerving to the right every time there's a merge lane.

This must be a difficult problem or have other unintended behaviors if fixed. Maybe they can have the B Pillar camera recognize that there is a merge lane there. I notice that NOA already starts showing and reacting to cars merging so the B pillar is doing this now, they just aren't using this as an input to avoid the lane splitting of the merge lane.

 

[eloder]

I'm very glad I ended up purchasing it. Did a test drive with EAP before delivery, and I was pretty amazed at how much it improved in just a year. Using it so far has been awesome--at the very minimum I feel like it's a nice safety feature as a "second set of eyes". For example at night I feel a lot more comfortable driving while I supervise on AP. Especially when blinded by extra bright headlights! It's even caught some things ever so slightly faster than I would have (drivers making wide turns across incoming traffic), AP slows down before I even hit the brake.

I think the important thing is just knowing what factors it cannot handle to avoid surprises.

It will get better over time, and I'm excited to see what this will look like in another couple years. I'd say my only big complaint right now is the lane change disengagements that happen so far into the procedure.

 

[mrrross]

I love it and use it daily for 95% of my ~80 mile commute in LA area. After about a month of using it i've figured out where I use it and when I don't. It's not perfect, but very helpful easily worth it in my opinion.

 

[tpoltron]

It’s interesting how much variation there is between people in their comfort trying and using AP. My wife is unnerved by it while I find it highly desirable as it is, though admit AP still has a way to go. It’s a godsend in start n stop traffic on jammed California ‘freeways’ and great on an open road.

Many of our highways are 3 or 4 lanes wide so I’ll often go in a middle lane to avoid the exit ramps swerves. Did a 230 mile drive yesterday in a mix of heavy traffic and windy mountain road and AP did great. It only freaked out once when the middle lane split in two to add a truck climbing lane and it didn’t know which one to take. Otherwise for 5 hours it was amazingly centered and smooth (for both steering and braking) with just a little steering roughness in the tightest couple mountain curves (and oh yes my hands were on the wheel) and some speed hesitation passing big trucks on curves (no worse than many human drivers).

 

[JeffK]

1. The "put pressure on the wheel" occurs about 3x/mile a real pia

1. I hold both hands on the wheel and don't rest my elbows on anything... the nag almost never appears.
2. Yes, that's how auto centering works.
3. I've not experienced this...
4. lane change too fast? I'd say lane changes are too slow for me.

Please don't forget EAP is considered an ADAS (advanced driver-assistance system), the goal is to assist you as the driver.

Do you think it accomplished this? If so, then it's ready for prime time.

Personally, I love it, I use it every day for my commute to work.

 

[ℬête Noire]

I regularly use it on local "side" roads. I could take it or leave it there. However for long trips I find it absolutely amazing.

1. The "put pressure on the wheel" occurs about 3x/mile a real pia

I never take my right hand off the wheel, holding it with 2-3 fingers at about the 4 o'clock position. So far I'm using 2 seat portions, "I'm driving" and "overwatch" (I don't literally use those tags, but that's the concept). Overwatch allows me to use the center console lid as an armrest while gripping the wheel. The grip is only hard enough to have a bit of friction on the wheel covering, and I am constantly pushing right or left slightly depending on which way I'd rather the car "fail" towards. For example on the outside of a turn I'm tugging slightly towards the center of the road.

When the road is dead straight and I have no preference this gets a bit weirder, as you have to go to some "random" direction. That's where I'll occasionally get "flashing blue" to remind me to put feedback into the wheel. That visual signal is so high up that it's easily in my field of view as I'm watching the road, so very rare for it to progress to the audio reminder. If it gets to the audio I have my doubts I was properly paying attention to the road, and so I find it a good reminder to re-focus.

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.

EAP knows no fear! Embrace the EAP! The start of a 3rd "passing" lane on 2-lane highways is another place it'll work hard to find the new lane. I used to flip out of EAP coming up on these (because I'm always ready to flip into manual mode), taking full manual control, but most of the time I don't even bother anymore and let the car sort it out. Not only because it's been improving on it's behavior to some degree, but because I'm used to it (6500 mi, a little over 2 months in). The key is in the name Overwatch. I'm not the least bit surprised when the car behaves like this, I know when it is going to have to work hard to find the "center" of a dubiously marked road, because I'm watching and see stuff coming up.

I also am getting close to the right spot where I don't need two seat/wheel profiles. I've been refining these and they are converging. They may join at some point soon.

P.S. I have found the new Lane Change on Nav is currently nigh useless for this area, though. I don't know if anyone is finding that works well?

 

[boaterva]

In case it’s not clear to OP, the nag means ‘keep the wheel from turning slightly’. ‘Pressure’ (squeezing) does nothing. I’ve requested they change the text to ‘apply turning pressure’ or something. But they probably think that’s too complex.

And the recent addition of using any steering wheel control to ‘answer’ also helps!

 

[Twiglett]

I think the OP has pointed out legitimate issues that the current version of AP has... However as someone who's been driving on AP 2.0 for over a year and a half, this is by far the most mature version and it's way better than anything else you can get in a consumer vehicle.

I do really wish that Tesla would find a solution to the issue with the car trying to center on merge lanes where the dotted lines go away as you are driving in the right lane. This is by far my biggest issue when using Autopilot on a 2 lane highway. I either need to be in the fast lane, or in the slow lane and tolerate the car swerving to the right every time there's a merge lane.

This must be a difficult problem or have other unintended behaviors if fixed. Maybe they can have the B Pillar camera recognize that there is a merge lane there. I notice that NOA already starts showing and reacting to cars merging so the B pillar is doing this now, they just aren't using this as an input to avoid the lane splitting of the merge lane.

this is my point - the OP has a personal opinion, to which he is perfectly entitled, but there are huge numbers of us that use it every single day and love it for what it is.
The problem with waiting for perfect - sorry "ready for prime time" - is that nobody would have it and we would be waiting for years until it was deemed perfect for the ignorant masses. Thats how the dinosaur burner car companies work.
I'd rather have it right now thanks and appreciate every single little improvement along the way.
DevOps development in action

 

[apsen]

P.S. I have found the new Lane Change on Nav is currently nigh useless for this area, though. I don't know if anyone is finding that works well?

It works ok for me though it definitely doesn't see as far ahead as I do so sometimes I need to ignore the lane change suggestions.

 

[JeffK]

I need to ignore the lane change suggestions.

This, yes. Especially in mad max mode. If a semi truck is to the right of me slightly ahead of the line of cars trying to pass it suggests I get behind it

It also suggested I get into an exit only lane nowhere near my exit and it's previously suggested I switch lanes to a lane that ended in a few hundred feet. It's pretty funny sometimes.