Sponsored

DJG

Well-Known Member
Joined
Apr 13, 2022
Threads
12
Messages
1,006
Reaction score
1,106
Location
TX
Vehicles
Various
You might be right. That section of the road does have several different textures but the lane markers were pretty clearly marked along the way. I can't be sure. All I know is that I've had about 400 miles of self-driving so far and it was otherwise pretty smooth. I find it especially useful in stop and go traffic conditions.
I think it would be helpful, for you in particular, but all of us generally, to see if Rivian could give a post mortem/debrief of sorts of what they think the system saw/was doing. It could help everyone anticipate weaknesses in their own use.
Sponsored

 
OP
OP
trez63

trez63

Well-Known Member
Joined
Mar 7, 2022
Threads
40
Messages
349
Reaction score
624
Location
SoCal
Vehicles
R1T, Gladiator, Porsche GT4, and a bunch of bikes
Clubs
 
this is a fairly "normal" event with autosteer systems.
:) This isn't the only car I've ever driven/owned with autosteer type functions. But it's the only one that's tried to smash me into the wall at 80mph. Lane centering is relatively mature now in most vehicles and if it's not ready then it shouldn't be two taps away from the average user.

It's not a self driving system
I get that. To some it's semantics and to others it's the difference between night and day. And perhaps the title of the post is the reason you bring this up. But lane centering, or autosteer, or whatever you want to call this feature is extremely dependable in many cars. The purpose of this post is to inform those of us who aren't aware, that it's not to be depended on in the R1T. You seem to already know that, but I am quite sure not everyone is as well informed and well researched as you are on the topic. I for one was not expecting it to do what it did and was quite shocked.
Again, what you see in the video doesn't seem nearly as aggressive as it was in real life.

One more note. The light conditions were not nearly as dark as my dashcam makes them out to be. It was quite a bit lighter out than it seems from the footage. Not sure if that matters, but I figured I'd add that.
 

zefram47

Well-Known Member
First Name
Aaron
Joined
Feb 6, 2022
Threads
13
Messages
2,019
Reaction score
3,159
Location
Denver, CO
Vehicles
R1T, C6 Corvette GS
Occupation
Software Engineer
Clubs
 
The purpose of this post is to inform those of us who aren't aware, that it's not to be depended on in the R1T.
It's not to be depended on in ANY vehicle that has a Level 2 ADAS system. But you ignored my other post, so I'll repeat it. Those that act as if any of these Level 2 systems can be treated as self-driving is why several of us keep bringing up the semantics. It matters when the average person thinks they actually are self-driving...they aren't.
 

Gator42

Well-Known Member
Joined
Dec 3, 2020
Threads
13
Messages
880
Reaction score
1,455
Location
SC MA NV
Vehicles
2020 Defender
Those that act as if any of these Level 2 systems can be treated as self-driving is why several of us keep bringing up the semantics. It matters when the average person thinks they actually are self-driving...they aren't.
Even the term ‘Level 2’ is irresponsible as it implies another, more capable level. It’s all hubris…
 

DJG

Well-Known Member
Joined
Apr 13, 2022
Threads
12
Messages
1,006
Reaction score
1,106
Location
TX
Vehicles
Various
:) This isn't the only car I've ever driven/owned with autosteer type functions. But it's the only one that's tried to smash me into the wall at 80mph. Lane centering is relatively mature now in most vehicles and if it's not ready then it shouldn't be two taps away from the average user.


I get that. To some it's semantics and to others it's the difference between night and day. And perhaps the title of the post is the reason you bring this up. But lane centering, or autosteer, or whatever you want to call this feature is extremely dependable in many cars. The purpose of this post is to inform those of us who aren't aware, that it's not to be depended on in the R1T. You seem to already know that, but I am quite sure not everyone is as well informed and well researched as you are on the topic. I for one was not expecting it to do what it did and was quite shocked.
Again, what you see in the video doesn't seem nearly as aggressive as it was in real life.

One more note. The light conditions were not nearly as dark as my dashcam makes them out to be. It was quite a bit lighter out than it seems from the footage. Not sure if that matters, but I figured I'd add that.
Sure, my only point is that this can happen in any vehicle with this system, and it has quite frequently for me (probably 1 out of every 5 or 6 long drives I used it). And that's in systems (eTron) that I'd considered advanced and that I'm fully confident in. The caveat is that fully confident still means knowing it can turn you into a wall at any moment (because it doesn't know the wall is there and might think something is in the road that is not).

Personally, I think being surprised by an event like this is a sign of being too trusting in a system's (any system) capabilities. And your statement that it can't be depended on in the R1T goes without question. It can't be depended on in any vehicle. If anything, the Rivian system should be trusted less than any other system on the road, because there is no objective and rational data/reason to trust it at this point (though subjectively I do find it to be very good relatively speaking in my limited use thus far).

This is also part of the reason why I think Rivian hasn't unlocked the hands free features yet. If something like this had happened under that system (and it definitely will) at this early stage, now that would be a story. They need to build up more strength and credibility before defending themselves in that situation. You can't just come out with hands free out of the gate, lawyers would have a field day.

Again, glad everyone is ok to be having the discussion, but by no means should it be viewed as a fluke event. It's a reminder that it can and will happen again.
 

Sponsored

OP
OP
trez63

trez63

Well-Known Member
Joined
Mar 7, 2022
Threads
40
Messages
349
Reaction score
624
Location
SoCal
Vehicles
R1T, Gladiator, Porsche GT4, and a bunch of bikes
Clubs
 
Sure, my only point is that this can happen in any vehicle with this system, and it has quite frequently for me (probably 1 out of every 5 or 6 long drives I used it). And that's in systems (eTron) that I'd considered advanced and that I'm fully confident in. The caveat is that fully confident still means knowing it can turn you into a wall at any moment (because it doesn't know the wall is there and might think something is in the road that is not).

Personally, I think being surprised by an event like this is a sign of being too trusting in a system's (any system) capabilities. And your statement that it can't be depended on in the R1T goes without question. It can't be depended on in any vehicle. If anything, the Rivian system should be trusted less than any other system on the road, because there is no objective and rational data/reason to trust it at this point (though subjectively I do find it to be very good relatively speaking in my limited use thus far).

This is also part of the reason why I think Rivian hasn't unlocked the hands free features yet. If something like this had happened under that system (and it definitely will) at this early stage, now that would be a story. They need to build up more strength and credibility before defending themselves in that situation. You can't just come out with hands free out of the gate, lawyers would have a field day.

Again, glad everyone is ok to be having the discussion, but by no means should it be viewed as a fluke event. It's a reminder that it can and will happen again.
Agreed on all points. I think you and I are saying the same thing. With the exception of the "surprise" of course. If I had known (as I do now) how unpredictable it is I would have never used it in the first place. And it begs the question whether something so unpredictable can really be considered safe for use.

I know the impulse to protect the brand. I have the same impulse. But when it smashes your kid's head on a calm drive back from a staycation you tend to take off the fanboy hat and get objective. I don't remember signing something that said I was going to field test an experimental feature. I do like how a few of the responses are blaming me for trusting the car more than they're blaming the car though. That totally checks out. (not you DJG)
 
OP
OP
trez63

trez63

Well-Known Member
Joined
Mar 7, 2022
Threads
40
Messages
349
Reaction score
624
Location
SoCal
Vehicles
R1T, Gladiator, Porsche GT4, and a bunch of bikes
Clubs
 
It's not to be depended on in ANY vehicle that has a Level 2 ADAS system. But you ignored my other post, so I'll repeat it. Those that act as if any of these Level 2 systems can be treated as self-driving is why several of us keep bringing up the semantics. It matters when the average person thinks they actually are self-driving...they aren't.
Do you want me to submit a formal apology for using that term? Go ahead and write it up in your own terms and I'll sign it.
Let's stop acting like everyone is supposed to know the full terminology and capability of every release level of ADAS. I don't want to read too much into your tone here, but it's starting to seem like you want to blame me for depending on the truck more than you want to blame the lane keeping capabilities of the truck. I will remind you that I was fully paying attention and not solving a crossword puzzle while the RIVIAN did all the driving. That's probably why I'm still here.

There's a bigger point here.
It's not that the RIVIAN failed to avoid a hazard in the road that I wasn't paying attention to. That would be my fault at this stage of the technology.
The RIVIAN swerved off a perfectly normal road towards the center divider despite my vigilance.
If you can't tell the difference between those two scenarios there is really no point to us trying to come to an understanding.
 

Engi_Nerd

Well-Known Member
Joined
Jul 21, 2021
Threads
6
Messages
220
Reaction score
425
Location
Western NY
Vehicles
Model Y LR, Alfa Romeo Giulia
Glad everyone is okay and that you were using it so responsibly. Near misses will be a regular occurrence as Rivian builds out their system, however that was a pretty bad one as they go. No vanilla lane keeping system should be inputting enough torque to hit passengers heads on the roof. I would definitely share it with them and request feedback. Not sure what's to gain by having it on the web. The reality is that these systems need to accumulate millions of miles to start finding edge cases like this and this is yet another area where early adopters are doing the heavy lifting.
 
OP
OP
trez63

trez63

Well-Known Member
Joined
Mar 7, 2022
Threads
40
Messages
349
Reaction score
624
Location
SoCal
Vehicles
R1T, Gladiator, Porsche GT4, and a bunch of bikes
Clubs
 
Not sure what's to gain by having it on the web.
If I had seen this video posted here I wouldn't have used the feature with my family in the car at night going highway speeds no matter how alert I was.
 

Sponsored

MountainBikeDude

Well-Known Member
First Name
Adam
Joined
Jun 20, 2019
Threads
40
Messages
1,878
Reaction score
3,845
Location
Vancouver
Vehicles
2023 El Cap Quad Motor R1T (Selling the Xterra)
Clubs
 
It could be it thought there was an object in the road. My adaptive cruise has been fooled by the shadow of an overpass once. Hard to see it in the video but upon watching a few times, it looks like the truck might be trying to avoid the darker section of pavement that doesn't quite take up the whole lane.

1656449296431.png

1656449529441.png
Good point and highly probable.
 

DJG

Well-Known Member
Joined
Apr 13, 2022
Threads
12
Messages
1,006
Reaction score
1,106
Location
TX
Vehicles
Various
Agreed on all points. I think you and I are saying the same thing. With the exception of the "surprise" of course. If I had known (as I do now) how unpredictable it is I would have never used it in the first place. And it begs the question whether something so unpredictable can really be considered safe for use.

I know the impulse to protect the brand. I have the same impulse. But when it smashes your kid's head on a calm drive back from a staycation you tend to take off the fanboy hat and get objective. I don't remember signing something that said I was going to field test an experimental feature. I do like how a few of the responses are blaming me for trusting the car more than they're blaming the car though. That totally checks out. (not you DJG)
I certainly am not blaming you for what happened in the least, there isn't any blame to be placed in my mind actually. None of this has anything do with you or even Rivian in fact. It's a somewhat predictable event (statistically speaking, not the exact time and place) when using a lane keeping system at this time, that's the point I keep making. If you had said you were actually driving a Tesla, or an Audi, or a XXXX, my reaction to the video and comments would all be the same - Yep, looks about right, I've had that happen before. I've had that happen at least a couple dozen times in two years with an Audi and at least once already with the R1T. In my Audi I know that when coming up to a light with left turn lanes and I'm in the current far left lane, there is a small chance it will want to jump into that newly created turn lane at the last minute with cars stopped at the light even though I want to continue through the green light at 65mph. I also know that the Audi system is atrociously bad at approaching cars stopped at a red light on highways. It's radar must be very limited range, because it won't pick up those cars until it's far too late and I'd slam into them at 60mph. My 10 year old Infiniti will handle the same situation flawlessly with it's simple adaptive cruise function. I now understand these nuances and apply them. I grip the wheel firmly in the Audi when approaching left turn lanes, and proactively apply my own brakes when approaching a red light.

It comes with the territory of using the system, and while we don't explicitly agree to the risks, we are implicitly agreeing to them by using it and only after using it do we fully understand the risks (both in general and specifically for that system). There is no regulatory agency (yet) that tests/certifies a system and allows an automaker to use it or advertise it however they so choose. It's a market regulated system still, and in a market regulated system, caveat emptor is king.

This is a discussion on driver assistance systems in 2022, not necessarily a discussion of Rivian's system. It would be irrational to render judgement of any one system based on a single event. Especially without even knowing what caused it to happen. Was it a fault in the system ("bad code"), or was it actually acting as it should but it was just a "perfect storm" of conditions that made it do something that in hindsight was not desired? Did it actually think it was saving your life by avoiding an obstacle in the road that came out of nowhere? The takeaway for everyone is different depending on the answer, which is why I'm interested to hear if Rivian weighs in on the details of this event.
 

Weeyam eyesore

Well-Known Member
Joined
Mar 12, 2022
Threads
8
Messages
46
Reaction score
36
Location
Central pA
Vehicles
Ford maverick
Occupation
Working
This shit happens to me often when I use driver +

I rarely use as a result. I do not find it helpful. Actually more dangerous currently
 

Gator42

Well-Known Member
Joined
Dec 3, 2020
Threads
13
Messages
880
Reaction score
1,455
Location
SC MA NV
Vehicles
2020 Defender
Are we all still here? 🥸 Feels like this thread could come up ‘missing’
Sponsored

 
 




Top