Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register

Consumer Reports: Tesla Autopilot a “distant second” to GM Super Cruise

This site may earn commission on affiliate links.
I’m referring to the linked video above with a review of the system and to how it functions. Not in regards to how it monitors. The actual driver assist abilities were less than impressive. Again, I’ve never ridden in one. Only basing this off that video.

Here's a better video. This is the newest SuperCruise with auto lane change. It does lane keeping and following the car in front on interstate highways similar to AP on highways. It also does auto lane change when you tap the stalk just like Teslas. And it is completely hands-free. And it works with sun glasses too. He is also wearing a mask and it does not affect the drive facing camera system. it still works.

 
You guys are missing the point...Tesla Autopilot was clearly deemed the best system in this CR review (it just has potential safety issues which resulted in a demerit to the "score" but that's not the point). But you're missing the one suggestion in the article which unquestionably would improve the Autopilot experience and make it MUCH easier to use, and better:
----
The system should also encourage drivers to provide their own steering inputs by keeping the system active rather than going into a standby mode,” says Funkhouser. Most of the systems we tested do allow drivers to steer when they want, perhaps to avoid a pothole or another car driving too close to the lane line. “If the system works well, it will gently nudge the car back to the center of the lane after the driver stops providing steering input,” Funkhouser says.


If you try to steer around a pothole in a Tesla, however, the auto-steering system will resist, unless the driver applies a significant amount of torque to the steering wheel, after which the system will abruptly shut off, leaving the driver in control.
We don’t like that Cadillac’s system goes into a standby mode as soon as the driver applies additional steering input, which also can dissuade the driver from taking over if needed. We also found that Cadillac’s system is finicky about re-engaging, even when the driver is well within the lane lines.
----

Tesla should definitely improve their system to allow blended inputs. I do wonder how much of the fact that they don't do this has to do with the torque-sensing requirement they have right now, which is inherently a driver input, with a dead zone. Dead zones are bad!

I'm hopeful that maybe they can get the cabin camera working well enough that at some point in the future they may relax the torque requirement, or modify it so that it blends out and enables this much better function, which would unquestionably be better than what we have right now in all ways (since it has everything we already have with no regression, but with more function). I'd love to be able to make microadjustments to steering without having to jerk out of AP, to avoid potholes, road debris, etc. It's pretty annoying to have to remember to disengage autopilot and then re-engage it to avoid jerking.

As the article states, it's also safer, as it encourages active driver participation & involvement in the driving task, rather than discouraging it.
 
Here's a better video. This is the newest SuperCruise with auto lane change. It does lane keeping and following the car in front on interstate highways similar to AP on highways. It also does auto lane change when you tap the stalk just like Teslas. And it is completely hands-free. And it works with sun glasses too. He is also wearing a mask and it does not affect the drive facing camera system. it still works.


Thank you. Appears better, but also being shown by a GM Super Cruise spokesman. But yes, certainly seems better, albeit at the level of AP 2 years ago and still very much constrained to their mapped highways. No thank you.
 
Super Cruise's driver facing camera can tell if the driver is dead drunk and safely disengage. That's the whole point of a driver facing camera. It can do that. The wheel move nag system cannot detect if the driver is dead drunk. That's why the driver facing camera is a better system.

If the driver is dead drunk, dead, or drunk, I’d rather it be engaged. At least then he has a fighting chance of surviving.

Er....except if he’s dead.
 
  • Like
Reactions: DanCar
Here's a better video. This is the newest SuperCruise with auto lane change. It does lane keeping and following the car in front on interstate highways similar to AP on highways. It also does auto lane change when you tap the stalk just like Teslas. And it is completely hands-free. And it works with sun glasses too. He is also wearing a mask and it does not affect the drive facing camera system. it still works.


It’s cool how it works hands free, just wish I could understand him better without that mask on. Sounded a bit muffled, wasn’t sure I could pick up on all the features GM was trying to convey? Maybe I’m just old? Maybe a GoPro next time?
 
  • Funny
Reactions: AlanSubie4Life
You have to wonder next year when CR review the Tesla with FSD what they will think. While I admire CR for attempting to be objective and avoid advertising-driven bias, they do tend to be somewhat stuffy in the way they approach things. You get the impression you are reading a review written by your grandmother .. it's all so comfortably numb.
 
  • Like
Reactions: Fernand
Except it does not do that. You can easily satisfy the nags without ever keeping a hand on the steering wheel. You can just roll the left scroll wheel with a finger anytime the screen flashes blue to make the nag go away. And of course, you can cheat the system completely with a anti-nag weight.

That's the whole problem with Tesla's torque system. It does not do a good job of making sure your hands are on the wheel. It is too easy to make the system happy without having your hands on the wheel or ever paying attention to the road.

We keep having this discussion. Is it really the cars job to make sure you are paying attention? Why? Sure, I know the argument, "With the car doing so much its easier to lose focus". Well that can apply to regular cruise control, car radios, automatic transmission, and even the heater (making you drowsy).

And where does that place liability? The more you do to ensure the driver stays alert, the more you are implicitly assuming responsibility. Apple have had to fight several lawsuits over crashes caused by people texting while driving; the claim was the iPhone should have detected the car was moving and disabled texting.

Personally, I find it obnoxious when people try to find someone else to blame for their own incompetence.
 
We keep having this discussion. Is it really the cars job to make sure you are paying attention? Why? Sure, I know the argument, "With the car doing so much its easier to lose focus". Well that can apply to regular cruise control, car radios, automatic transmission, and even the heater (making you drowsy).

I think the idea is to make sure deaths are lower with the system deployed than without it. That’s the goal! Distracted driving is a problem.

The more you do to ensure the driver stays alert, the more you are implicitly assuming responsibility.

Not a lawyer, but not sure whether straight-up negligence is a defense here. I think that Tesla and other auto makers will always be subject to automation related litigation.
 
Not sure how the report can be called unbiased when Tesla won capability and performance, and ease of use, but comes in a “distant second.” The categories are not weighted and most categories were clearly designed to play to the only perceived strength of SuperCruise. Who paid for the study?
 
We keep having this discussion. Is it really the cars job to make sure you are paying attention?

In a L2 and L3 system, yes. That's because the systems cannot work properly if the driver is not paying attention. The systems are dependent on the driver paying attention.

Systems like basic AP, NOA, and SuperCruise are designed to only do some of the driving tasks in a given set of conditions. This means that they NEED someone else (aka the driver) to monitor and do the other tasks when necessary. So you NEED a way to make sure the driver is able to do those other tasks that the car can't do. You can't have a system where the car only does say 90% of the driving and you have no way of knowing if the driver is able to do the other 10%, especially when that 10% might be critical for safety!

So making sure the driver is paying attention is fundamentally part of what makes the system work. If you don't have a driver attention system, you have a system that needs the driver to help with no way of knowing if the driver is able to help. Do you not see the problem with that?

Even Tesla has a driver monitoring system. Tesla did not deploy AP with no driver monitoring system at all and just hope that the driver pays attention. The only question is whether Tesla's driver monitoring system is the best system or not.

On a side note: that's why there is so much focus on autonomous driving. Many argue that a system that is dependent on the driver paying attention in order to work properly, is a fundamentally flawed system. They argue it is better to develop a system where the car is not dependent on the driver at all. If the car is able to do all the driving on its own, without any human supervision, then driver attention is irrelevant.
 
Last edited:
  • Disagree
Reactions: mikes_fsd
In a L2 and L3 system, yes. That's because the systems cannot work properly if the driver is not paying attention.

Systems like basic AP, NOA, and SuperCruise are designed to only do some of the driving tasks in a given set of conditions. This means that they NEED someone else (aka the driver) to monitor and do the other tasks when necessary. So you NEED a way to make sure the driver is able to do those other tasks that the car can't do. You can't have a system where the car only does say 90% of the driving and you have no way of knowing if the driver is able to do the other 10%, especially when that 10% might be critical for safety!

So making sure the driver is paying attention is fundamentally part of what makes the system work. If you don't have a driver attention system, you have a system that needs the driver to help with no way of knowing if the driver is able to help. Do you not see the problem with that?

Even Tesla has a driver monitoring system. Tesla did not deploy AP with no driver monitoring system at all and just hope that the driver pays attention. The only question is whether Tesla's driver monitoring system is the best system or not.

No, the question is how heavily you should weight the rating of a driver assistance system on the actual ability of the system to assist the driver, as opposed to its ability to handle the driver not paying attention.
 
No, the question is how heavily you should weight the rating of a driver assistance system on the actual ability of the system to assist the driver, as opposed to its ability to handle the driver not paying attention.

I would say both are important. And no matter how capable the driver assist system actually is at "driving" the car, you still need a robust driver attention system. You don't need less of a driver attention system if the driver assist is more capable. The only time you need less (or no driver attention) is if the system is no longer a driver assist but instead is an autonomous driving system. As long as the system is a driver assist, you need 100% driver attention all the time.
 
  • Disagree
Reactions: mikes_fsd
I would say both are important. And no matter how capable the driver assist system actually is at "driving" the car, you still need a robust driver attention system. You don't need less of a driver attention system if the driver assist is more capable. The only time you need less (or no driver attention) is if the system is no longer a driver assist but instead is an autonomous driving system. As long as the system is a driver assist, you need 100% driver attention all the time.

The reason why CR's reviews haven't been heavily weighting towards vehicles with driver monitoring systems in the past is that nobody gets distracted or falls asleep in a car without automatic steering.
 
The reason why CR's reviews haven't been heavily weighting towards vehicles with driver monitoring systems in the past is that nobody gets distracted or falls asleep in a car without automatic steering.

What!? Are you saying nobody ever gets distracted or falls asleep while driving their car? LOL.

And CR did not weight driver attention systems in cars with no automatic steering because the driver is in control of steering. The driver might get distracted or fall asleep but it is while the driver was in control of the car.

With automatic steering, it's different. The car is now controlling the steering, so the driver does not need to control the car anymore. So you do need to make sure the driver is paying attention.
 
Last edited:
What!? Are you saying nobody ever gets distracted or falls asleep while driving their car? LOL.

And CR did not weight driver attention systems in cars with no automatic steering because the driver is in control of steering. The driver might get distracted or fall asleep but it is while the driver was in control of the car.

With automatic steering, it's different. The car is now controlling the steering, so the driver does not need to control the car anymore. So you do need to make sure the driver is paying attention.

I was being sarcastic. I'm sorry. Sometimes I slip back into British mode where we assume people have a sense of humor and intelligence.

Firstly, they are treating Autopilot as a _safety system_, when the reality is that people want it as a convenience feature.

Second, CR has not previously used the same heavy weighting of driver monitoring in their ratings of vehicles, despite the fact that driver distraction and fatigue are _already_ significant safety problems.

Third, by their dumb scoring system, a perfect driver assistance system with no driver monitoring would score lower than an imperfect system with good monitoring. That is not rational.

This is a campaign, not a review.
 
So making sure the driver is paying attention is fundamentally part of what makes the system work. If you don't have a driver attention system, you have a system that needs the driver to help with no way of knowing if the driver is able to help. Do you not see the problem with that?

You could make the exact same argument for a vehicle that has no driver assist — L0 — and you would come to the same conclusion that there should be some system in place to be sure the driver is performing 100% of the duties required.
 
  • Like
Reactions: PACEMD