It was still pretty frequent up until the point where I moved from HW2.5 to HW3, which was... maybe October? Some time in the last half of last year, anyway. I doubt it has gotten better on HW2.5.
But yeah, it definitely seems better with HW3.
-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
GM Supercruise's 0 incident in 5.7 Million miles Vs Tesla Autopilot's 1 accident every 3.45 million miles (Safety Record Comparison)
- Thread starter Bladerskb
- Start date
It was still pretty frequent up until the point where I moved from HW2.5 to HW3, which was... maybe October? Some time in the last half of last year, anyway. I doubt it has gotten better on HW2.5.
But yeah, it definitely seems better with HW3.
rush6410
Member
You make a great point that irritates me. Most ppl can't tell you if braking was truly phantom or the car was responding to a condition that the driver is unaware of. I would suspect the later. Every "phantom braking" that I have encountered was actually the car responding to a condition. I always try to immediately identify why it applied the brakes so I can avoid the scenario which caused it. Most of the time it is a vehicle getting too close to the lane line next to you. Or, it perceives the vehicle is encroaching. But, I have yet come across a scenario in which I would consider it braking for no apparent reason. One funny or high stress scenario was when my car slammed on it's brakes while I was following another vehicle. My wife immediately was upset and told me to turn the thing off. I am sure a lot of ppl who can't identify the reason for the braking would respond the same way. What my wife didn't see was a car slammed on their brakes in front of the car I was following. The car in front of me swerved to the right to miss the stopping car. Since my car already was responding to the scenario, we safely avoided a major accident. My wife was so upset, it took a minute or so for her to calm down enough for me then to explain that the car just saved us. So, it is all about perspective. Some would respond like my wife and swear that the car tried to kill you. Then there is the real scenario where the car actually saved you.
rush6410
Member
What GM should really be asking is how many incidents occurred because they don't allow their driver to use SuperCruise outside of the geofenced areas which is 95% of the drivers trips. So basically they are ok with a higher rate of accidents as long as the driver was not utilizing SuperCruise. Sounds like a liability claim avoidance rather than what is best for the driver.
So, you’re saying your Tesla brakes when other cars get close to the lane line? I would be frustrated as well with that kind of behavior. Random slow downs on the street/freeway?
rush6410
Member
The adjacent car typically is crossing, touching, or approaching the lane line. Personally, I think it handles the encroachment pretty well.
powertoold
Active Member
I've yet to see an unedited SuperCruise video in normal traffic conditions where it performs well without randomly giving up with no audible warning (steering wheel just turns red and then gives control over to driver). Also, SuperCruise seems to take much longer to activate, unlike AP, which takes 1-2 seconds.
Check it out at 10:42 to see how finicky it is.
Check out 18:42 where it turns off for no reason.
SuperCruise does not instill confidence. The lane keeping is wishy washy and it only seems to work well in low traffic straight-ish road situations.
If you guys find a good video of merging traffic and bigger curves, do show.
Check it out at 10:42 to see how finicky it is.
Check out 18:42 where it turns off for no reason.
SuperCruise does not instill confidence. The lane keeping is wishy washy and it only seems to work well in low traffic straight-ish road situations.
If you guys find a good video of merging traffic and bigger curves, do show.
Last edited:
MXLRplus
Active Member
Drive West on the 91 fwy between 6am-11am on a weekday in the FasTrak. It happens like clockwork at the Riverside/OC county line, then again on the 91/55 transition. That's a ~15 mile drive. I have performed that drive in both a Tesla and an I-Pace w/autosteering hundreds of times. As of this week's update on HW3 in a 2020 MX the oh-shiits-per-mile is still too high to use AP at these areas. Bridge shadows remain a problem as well.
The good news is that people on this site make far more errors when they post than Autopilot™ does, or roughly one foot-in-mouth per 10 posts.
powertoold
Active Member
You said "on average every 25 minutes", insinuating an overall poor performance and then proceed to point out some particular road or interchange? That's cool.
I recently went on a road trip to around the OC area, 600 miles round trip, not a single phantom brake. It has improved lately.
MXLRplus
Active Member
I have more than 15k miles of freeway driving all over SoCal. The MX is my 'shop truck' replacing our pickup. While I quoted two examples that you can test yourself, there are dozens of areas in SoCal that will trigger AEB inadvertently. But it also triggers at random when there is a speed differential in congested situations, or in bright sunlight combined with dark shadows.
As of today, I still cannot have AP or FSD engaged with my wife in the Tesla in traffic. Note that her CT6 also has ACC/AEB as do two of our other cars. We are no strangers to advanced driver's aids, and we are both professionally trained competition drivers.
BTW - What motive do you perceive that I have that would make me misrepresent a road test review?
Krugerrand
Meow
Don’t be obtuse. Of course the car slows when a vehicle beside starts to encroach into its lane. You’re not one of those people who drive along stubbornly thinking, ‘I own this lane. If you hit me, I’m suing you.’ Of course you’re not. You’re someone who would yield to an encroaching car to avoid a potential accident because that’s what reasonable, intelligent people do - and so does a Tesla on AP.
MXLRplus
Active Member
Actually, it will do it at random (fairly rare) when the car you are passing is dead center in it's lane. Trailers seem to fail more often than sedans.
Soda Popinski
Member
Interesting. Just to clarify, a GM car on SuperCruise, not a Cruise autonomous car, right?
MXLRplus
Active Member
That Cruise collision in SF was 100% the squirrel's fault. The squirrel hit the car, not the other way around. He was one of those nitwits who splits any lane, any time, yet lacks talent. GM paid a nuisance fee.
Well in this case it sounds like his head mounted camera was causing this. So I wouldn’t fault it here. Your other time stamp however....
Barklikeadog
Active Member
Off Topic, how does a model x replace a pickup? Did you modify the vehicle? Remove seats?
You sound as if you're a lawyer defending Tesla's Autopilot.
If there's a car driving along besides me and it comes closer to the lane marker between my lane and theirs, I may move over slightly to the other side of my lane. But I do not slow down or step on the brake. You do? That's strange.
This is what the previous poster was talking about, that his Tesla is applying the brakes in this situation.
If slowly drifting over up to the line, then yes I do what you do. But if it looks like the person will continue moving over into my lane, I do reduce speed or get out of there. I do think AP slowing down is a good thing but just not quite so quickly.
sorry. I talked to him and he said it was autonomous. He said just google GM autonomous hit motorcyclist and it pops up.
Similar threads
