AP1 vs. AP2

[Electroman]

The fact that AP2 recognizes cars and initiates the slow down from a greater distance than AP1 seems like big plus to me.

There are still many more pluses in the AP1 column for me.

I don't disagree with that.

But IMHO in general the differences are not significant enough to lose sleep over. If you read the posts of a few here, you would think as soon as you turn on AP the car would search and head towards the closest tree or ditch. And it looks like they have the core Autosteer/TACC functionality quite well completed with a few further refinements needed. Perfectly usable.

 

[studiojon]

I drive an AP1 car - have put 30k miles on in a little over a year. I do the same commute and have a good understanding of how it behaves in different situations. On current software it is really really good. More importantly, situations that might be problematic are easy to predict and define...

I had a loaner AP2 for an extended test drive to evaluate and see if I wanted to change. My answer - emphatically not yet.

AP2 was very good WHEN it was good. However the real issue was that it was much more difficult to predict how it would behave in different situations. there is much more uncertainty as to the path (road still jiggles) and you might get random steering for no good reason. I also believe it will get there and exceed AP1 but that we are still not at parity yet.

Additionally I have to say that the "silly smooth" algorithm also made a significant improvement to my AP1 car in terms of TACC handling.

 

[Electroman]

Additionally I have to say that the "silly smooth" algorithm also made a significant improvement to my AP1 car in terms of TACC handling.

That is true. In a lot of ways AP1 is like the gold standard, that you would be hard-pressed to improve upon in highway driving. It just simply drives !!

 

[OilSucks]

Tesla pretty much misled everyone who bought before December 2016 about the safety features (someone has posted the wording but it was definite on the safety features). EAP and FSD had various limiting language, for EAP it said regulatory approval and that it would begin to be rolled out via software updates in December 2016 but none of the safety features had that language.

So I don't think its overly dramatic to say Tesla lied to all buyers about the safety features that would be present on HW2 vehicles. AEB and these features have a real $ value and I likely would've waited or bought AP1 at a significant discount had I known Tesla wouldn't have AEB ready for 6 months! In fact, all the sales people and my DS, in writing, said, it would have the safety features and AP1 parity by end of the year (when I got my car). I ordered in late October/confirmed early November.

Same, I would have happily picked up an AP1 Model X for around the same cost as my brand new AP2 Model S. What kills me is that I was so green and new to the Tesla scene I didn't realize I was test driving an AP1 Model S, got so excited for it and purchased the EAP/FSD package, only to discover that I bought an AP2 Model S.

 

[BLKTSLA]

Did a side-by-side to see the differences as of the date of this posting.

 

[timvracer]

Conclusion:

You see more things on the display with AP1.0 if you like graphics. Otherwise 2.0 is at parity or better in actual driving (unless you'd rather sometimes drift right instead of drift left).

Did a side-by-side to see the differences as of the date of this posting.

 

[BLKTSLA]

Conclusion:

You see more things on the display with AP1.0 if you like graphics. Otherwise 2.0 is at parity or better in actual driving (unless you'd rather sometimes drift right instead of drift left).

Not quite yet, there is the TACC disparity which will be on the night version of the comparison. Not to mention Auto-wipers which I don't attribute to AP functionality but are a byproduct of the hardware. I'd say AP2 is 85% to parity, problem is it should be further along at this point. If it is truly at parity then we should start to see the "Enhanced" features starting to roll out, would love to get all thats mentioned here:

 

[u00mem9]

Conclusion:

You see more things on the display with AP1.0 if you like graphics. Otherwise 2.0 is at parity or better in actual driving (unless you'd rather sometimes drift right instead of drift left).

I’ve always wondered if the higher ride height and camera position if the X allow better AP operation especially over crests in the road.

 

[EV-lutioin]

I have AP2 in my Model 3 and on my commute it is limited to 45 mph on a 55 mph road, but my Model X with API is limited to 60. Anyone else experience this?

 

[TaoJones]

I have AP2 in my Model 3 and on my commute it is limited to 45 mph on a 55 mph road, but my Model X with API is limited to 60. Anyone else experience this?

Sounds like the culprit is the buggy AP2 database(s) associated with both speed limit values and road types. If either or both are incorrect, the scenario you've described can happen with AP2, while you're fine with AP1.

It's particularly bad in California and in Alberta, I've noticed.

 

[BLKTSLA]

 

[Electroman]

can't tell the difference..

 

[aikisteve]

I also did a side by side review a few weeks back:

 

[rotaz]

So what's your verdict?

 

[BLKTSLA]

So what's your verdict?

My verdict is that AP1 is still superior with the edge going to AP2 for lane keeping and recognition.

AP1 can recognize speed limit signs, identify other types of vehicles on the road (trucks, motorcycles). It also does a better job of dealing with a single line on the road by hugging that 1 line until the parallel line comes back.

AP2 is superior in seeing the lanes (no fading in and out) and keeping the lane. Due to the wide angle camera, so it can see wider lanes where AP1 cannot. However, when it has just 1 lane line, it wonders/swerves searching for another line instead of just keeping the 1 line. AP2 also displays more cars on the road than AP1 but its nominal since it has the same field of view (by software) as AP1.

 

[lunitiks]

My verdict is that AP1 is still superior with the edge going to AP2 for lane keeping and recognition.

AP1 can recognize speed limit signs, identify other types of vehicles on the road (trucks, motorcycles). It also does a better job of dealing with a single line on the road by hugging that 1 line until the parallel line comes back.

AP2 is superior in seeing the lanes (no fading in and out) and keeping the lane. Due to the wide angle camera, so it can see wider lanes where AP1 cannot. However, when it has just 1 lane line, it wonders/swerves searching for another line instead of just keeping the 1 line. AP2 also displays more cars on the road than AP1 but its nominal since it has the same field of view (by software) as AP1.

^ This

 

[BATT_MBL]

So are the differences in AP1 and AP2 still minimal at this time (Aug 2018)?

 

[aikisteve]

So are the differences in AP1 and AP2 still minimal at this time (Aug 2018)?

With 2018.28.1 AP2 has regressed quite a lot. I've just put up a post about it with a video of my test. So, AP1 is now even more a clear winner. the promised v9.0 will have to work miracles to be better than AP1 now.

 

[bhzmark]

AP2 beats AP1 on every dimension tested: Tests uncover issues for advanced features

AP1 especially struggling on hills:

 

[bmah]

AP2 beats AP1 on every dimension tested: Tests uncover issues for advanced features

From the article:

"The 2017 BMW 5-series with "Driving Assistant Plus," 2017 Mercedes-Benz E-Class with "Drive Pilot," 2018 Tesla Model 3 and 2016 Model S with "Autopilot" (software versions 8.1 and 7.1, respectively) and 2018 Volvo S90 with "Pilot Assist" were evaluated."

The AP1 (2016 Model S) and AP2 (2018 Model 3) cars weren't running the same version of code, so you can't really make the blanket statement "AP2 beats AP1" based on that evidence.

(I personally am not claiming that either AP1 or AP2 is better by any particular measure. I've only driven AP1 and non-AP cars.)

Bruce.