Autopilot Tops Super Cruise Capability: Dinged For Driver Engagement

Tesla Model 3


Victim of its success.

When it comes to advanced driving automation systems, most people might think that Tesla is king. Autopilot is, after all, one of the most popular features available in the Californian electric vehicles, and we’ve seen it tackle some pretty challenging stretches of road. Well, maybe not. In a comparison test of a number of systems by Consumer Reports, Cadillac’s Super Cruise actually came out on top. And this isn’t the first comparison of these two systems have reached that same conclusion. So, what is going on here? Let’s take a closer look.

In everyday usage, there’s no doubt that many would prefer Autopilot, especially if they are Tesla owners familiar with its proficiency at reading the roads and adjusting to traffic accordingly. And, despite it being referred to as a beta system and recommended for highway use only, it can be engaged anywhere. Did we mention that the system is constantly being improved upon and given even more functionality over time?

Super Cruise, on the other hand, actually scores lower than Autopilot when it comes to basic functionality. It’s just not as capable and it’s also geo-fenced, meaning that it can only be used on certain highways. When it first rolled out, the number of roads where it would even turn on as quite limited, but that’s has grown over time and it now functions on a wide swath of our highway system.

So how does the Cadillac system manage to beat out Tesla’s? It all comes down to safety and it seems here, Autopilot is a victim of its own success. Experts in the field contend that with increased confidence in a system, comes a greater liklihood a driver will be less attentive. Sure, the Tesla system does nag you visually and audibly if you stay hands-off for 20 seconds or so, but a car can cover a lot of ground in that time and, perhaps, find itself in a situation where it needs the driver to intervene.

By contrast, the Super Cruise system employs a camera trained on the driver and monitors their eyes to see if they look away from the road. Close your eyes or turn your head for more than four seconds and the Caddy’s steering wheel lights up and it gives an audible warning. The seat may even vibrate.

While Autopilot is expected to get even more improvements and features in an upcoming “Version 9” update, it still won’t have this sort of advanced driver monitoring, and so we expect it will continue to lag behind Super Cruise in these sorts of comparisons for some time. The Tesla Model 3 does have an interior camera that possibly could be adapted to this purpose, among other things, but there’s no indication that will happen. When CR contacted the company about whether its interior camera might be purposed with this driver monitoring task in the future, they were told by email the following:

It might be utilized in potential future features, which could be added to Model 3 with software releases. Customers will receive prior notice if/when Tesla decides to use it.


Source: Consumer Reports

Categories: Comparison, Tesla

Tags: , ,

Leave a Reply

55 Comments on "Autopilot Tops Super Cruise Capability: Dinged For Driver Engagement"

newest oldest most voted

Good thing they got this press release out in time to beat Tesla System 9. Now we at least will have a base for comparison. While there might be some value to adding Lidar to a subset of Tesla’s to assist with more completely mapping or static imaging the surroundings, The Bosch radar imaging and cleverer software is already mapping every road worldwide that EVERY Tesla with Autopilot 2.0 or greater in real time which makes the static LIDAR maps of the Caddy seem very limited. So… how often does GM promise to update SuperCruise’s backing data and maps. I fear it will be like the now 4 year old maps in my 2014 Nissan. So… we’ll see… but I’m not sorry I drive a Model X.

Well the number one advantage of AutoPilot 2.0 and greater is that EVERY Tesla vehicle is equipped with it compared to the very limited number with SuperCruise.

As Jeffrey just wrote, Autopilot is constantly mapping everywhere Tesla cars drive and the AutoPilot system is constantly being improved and updated OTA so it only gets better and more capable.

No other manufacturer as yet does these things but I have a feeling that due to the unrelenting pressure from Tesla’s out-innovating them they will be forced to considerably up their game here and with electrification in general.

Tesla has already documented that AutoPilot reduces accidents and by reducing accidents it is saving lives and property.

But they have not released their protocol, if autopilot gets a driver into trouble and manual interruption fails to save their life.

I love AP, but not sure about Tesla’s stats.

Teslas have cameras and radar and is more capable than the system it’s being compared to. LIDAR doesn’t seem to be the issue here. It was something about the car not nagging you hard enough.

From what we’ve seen, vision is more than capable of figuring out what is and isn’t a road and LIDAR doesn’t help with lanes….

It seems lidar isn’t necessary only because current Tesla cars aren’t trying to do real-time mapping of the environment; no SLAM* system. That’s why they can’t detect stationary obstacles when moving at highway speed, and that’s why Tesla cars keep running into fire trucks parked on the highway. Yes, Tesla is constantly updating its geographical database, but that can’t possibly show temporary changes such as parked emergency vehicles or short-term detours due to road construction.

That’s not at all to say that you shouldn’t use Autopilot+AutoSteer if you are driving a Tesla car; the statistics quite clearly indicate that you are safer. But the current state of Autopilot in Tesla’s production cars is only Level 2+ autonomy, and what auto makers (and Waymo) are working toward is Level 4.

No auto maker, including Tesla, will ever get to Level 4 without a reliable SLAM, which will require active high-resolution scanners such as lidar or phased-array high-res radar. Elon Musk has his head buried in the sand denying that reality.

*SLAM stands for Simultaneous Localization And Mapping technology, a process whereby a robot or a device can create a 3D map of its surroundings, and orient itself properly within this map in real time.

So basically Autopilot works better and is easier to use, but doesn’t nag you as much as Super Cruise. I think Consumer Reports is good for a lot of things, but I don’t like some of their categories.

They get on missions to force companies to implement features (like safe window switches, lawn mower cut off switches, etc). Now, they appear to be doing that for driver’s assistants: encourage driver engagement.

It’s always been a problem with CR. They overemphasize certain things, particularly around safety, and downplay the things that really set cars apart and make them desirable. Go back 30 years and look at all the cars they rate highest and I bet you wouldn’t want to own the vast majority of them.

Yes, “reliability” (by which they mean merely the number of times a car is taken in for service) and safety seem to be almost the only things that concern Consumer Reports in their rankings for cars. Utility, comfort, luxury features, and whether or not a car is a pleasure or a pain to drive.. these things either don’t rate high on their list, or don’t appear on their list at all.

A glaring example of how CR’s criteria are seriously out of touch with what most car buyers actually think is important, is that the Tesla Model S consistently rates #1 on CR’s own annual survey for customer satisfaction, yet CR has repeatedly flip-flopped over the years on whether or not it will even recommend buying the car at all! 🙄

” yet CR has repeatedly flip-flopped over the years on whether or not it will even recommend buying the car at all! ”

CR’s recommendation is clearly stated on its website. It is based on multiple factors. One of them is reliability survey results. Regardless how good a car has done in other area such as buyer satisfaction or their testing score, it needs to score higher than average in order to be recommended. If a car’s reliability survey result comes in below that, it will lose CR’s recommendation, if it improves to be above it, then it will recommend it if other area are still stellar. That is why it may “flip flop” due to the yearly survey result of its owner’s reliability survey.

People on the comment section thinks CR flips flops because they are confused between the test score (highest ever), reliability survey (improving for Model S and still terrible for Model S) and owner’s satisfaction survey (consistently high for Model S and relatively high for Model X).

CR changed it’s rating for the Model S 3 times in a matter of months, with no new yearly survey results between rating changes.

CR should focus on the things that you as a shopper cannot do yourself. Everyone can judge for themselves how utilitarian, luxurious, or comfortable a car is.

Three out of the five categories are essentially “does the car physically prevent you from using this or feeling safe?” Also, what is a vertical line in a yellow circle supposed to mean? I’m assuming neutral?

Headline doesn’t seem to match text.

They’re saying the same thing, really, but in a different order (I didn’t write the headline).

It’s a little disengenious then. CR overall scores Supercruise as superior to AP, with AP winning some categories. At first read, sounds like the headline is saying AP, gets the overall win.

Yup “Autopilot Tops Super Cruise Capability: Dinged For Driver Engagement” should be “Super Cruise Tops Autopilot Capability: Autopilot Dinged For Driver Engagement”

Autopilot Tops Super Cruise ‘Capability’ – it wins on capability and performance, but that is negated by the driver engagement side

Tesla AP should absolutely be geo-fenced and not allowed to be engaged except on supported roads.

Since the focus of ADAS systems is primarily the safety of the driver and surrounding traffic, I can see why Super Cruise was deemed to have the ‘best’ system for the time being. The human brain is wired to basically be as energy efficient (read: lazy) as possible, so without strong driver engagement features, occupants will be relatively more prone to incidents than systems that basically require you to pay attention at all times. Both systems likely result in a significant reduction in crash risk that cars not equipped with ADAS face. People aren’t paying attention while driving everyday, hence why human error is estimated to be the largest contributing factor to roadway fatalities (~1.3 million a year) and injuries. No matter how many times Tesla says ‘Autopilot’ isn’t a Level 4 autonomous driving system, people are going to personally weigh the pros and cons and decide whether or not to engage the system in a given situation. GM simply took a more risk averse approach, Super Cruise was designed to provide so many checks on the driver that it becomes extremely difficult to attribute any blame to the system in the event of a collision. Likely after seeing the… Read more »

Well, the Tesla system has killed people and hit multiple fire trucks. The Cadillac system has not. The Tesla system is so lenient that it allows the driver to climb out of the driver’s seat and car proceeds on. It seems like a fair review. Now Cadillac just needs to work on the used interface I guess.

The Cadillac Super Cruise system has hardly been on a real road until now… and now only barely. We have yet to see how it handles when crazy folks try to hang a mask in front of the camera to fool it… just saying. Out of the hundreds of thousands of Tesla Autopilots driving down thousands of real world roads as I write this a tiny percentage have had serious accidents. Some of them caused by inattentive drivers or worse folks trying to fool the system or make it magically do more than it is currently capable of. I have owned an AutoPilot 2+ auto now for a year… and if I’m on a road it is likely on at least some of the time and on Freeways most of the time. It has actually stopped for me once when I missed one of the myriad hazards that appear everyday on the highways. I love it. When I first tried it the system was lousy at cresting hills… It used to panic when aimed at a concrete wall… because it was ignoring the curve in the road… now… these days with Over The Air updates my car is substantially better… Read more »

The “Tesla system” hasn’t killed anyone nor run into firetrucks. Drivers that didn’t operate their vehicles properly crashed. No ADAS replaces the driver’s responsibility.

Also, you can’t “climb out of the driver’s seat” and proceed. The seat has a sensor in it and Autopilot will disengage.

All true. But there have been at least a couple of “ingenious fools” who posted “Look Ma, no hands!” videos to YouTube, showing cases where the fools managed to defeat Tesla’s safety systems which prevent AutoSteer from being activated if there’s nobody sitting in the driver’s seat. Fools who defeated those safety systems by, apparently among other things, resting a heavy weight in the driver’s seat.

Naturally, anti-Tesla FÜDsters are using those crazy antics to claim Autopilot/AutoSteer is “unsafe”, altho of course that says a lot more about those FÜDsters than it does about Tesla Autopilot!

“It is impossible to make anything foolproof because fools are so ingenious.” — Murphy’s Laws

“Well, the Tesla system has killed people and hit multiple fire trucks.”

Typical Tesla bashing FÜD from a serial Tesla basher.

Tesla Autopilot has not killed anyone. What it has done is failed to save a few people who were not paying proper attention to driving.

But the number of lives Tesla Autopilot+Autosteer has saved, is far greater than the very few it has failed to save, as the accident statistics clearly indicate.

Capability is not about user interface

Isn’t the title backwards? “Autopilot Tops Super Cruise Capability: Dinged For Driver Engagement” but the article goes onto say that Super Cruise is viewed as superior to Autopilot by CR: “In a comparison test of a number of systems by Consumer Reports, Cadillac’s Super Cruise actually came out on top.”

Seems like the title should be “Super Cruise Tops Autopilot, Autopilot Dinged for Driver Engagement”

I don’t want a camera watching me. Sorry. If I want to get laid in my car for spontaneous sexy time I don’t want Tesla looking in

Model 3 has a camera pointed at the interior (including driver) too.

Having a camera pointed at you doesn’t mean that the video is being streamed and someone is watching it.

Besides the camera is pointed at your head if you are the driver. It won’t see your lap area.

No comment.

“If I want to get laid in my car for spontaneous sexy time I don’t want Tesla looking in”

Then I would suggest bringing someone else with you next time.

Get a room – And how about paying attention to driving while on the road…

They make tape for that.

Lol. Who uses tape these days. Mechanical media gave way to chips long ago. Even my home security uses SSD and cloud.

tape? haha

you know there’s tape and tape.

We’ll see how the marketplace responds, anything else is merely opinion.

I’d like to have the Tesla AP system with the Supercruise driver monitoring. That was one thing I found annoying about AP; constantly having to nudge the wheel to let it know I was still paying attention. Personally, I’d rather have a camera looking at my eyes to see I’m watching the road.

(⌐■_■) Trollnonymous

” Personally, I’d rather have a camera looking at my eyes to see I’m watching the road.”

How would Supercruise handle dark sunglasses?

It would think you are Supercool.

Probably uses infrared just like many iris cameras.

Just fine apparently from some of the reviews.

I think you should get nagged less, the slower the car is going.

Well, one is beta system so it allows driver to do more than they “should” so you can act like a beta tester to improve the system.

The other system is designed as a “final system” which may or may not be outdated but it is strictly controlled and went thru full Q/A to define a specified user parameter which tends to be more refined for those specified conditions but doesn’t allow you to use beyond its “intended” conditions.

Sounds just a difference in design approach.

CR should have done a test to compare the situation of an stationary object is on hwy, which system would have avoided it better. I want to know!

Tesla tried it (face camera) and decided that it was not reliable enough. CR didn’t live with driving the caddy system long enough to experience all of the false alarms, that’s all it means.

My personal choice? The car lights a lighted button on the dash, with a buzzer sound. You press the button to reset it. Then you get a cookie.

The system has been tried on monkeys for decades, and it works.

Shock therapy!

“Tesla tried it (face camera) and decided that it was not reliable enough. CR didn’t live with driving the caddy system long enough to experience all of the false alarms, that’s all it means.”

That is not the conclusion I’ve reached after reading multiple reviews of Cadillac’s Super Cruise. Just because Tesla couldn’t make the system work reliably doesn’t mean GM couldn’t find someone capable of developing that. Tesla’s software is significantly advanced in many areas, but that doesn’t mean their software engineers are better than everybody else’s in every single application.

I’m a strong Tesla fan, but that doesn’t stop me from recognizing that Cadillac’s system for detecting driver alertness is much superior to Tesla’s. This is one rare case where Tesla needs to follow, not lead.

Has anyone else had the reaction that CR publishing this story is like waving a red cape in front of Elon? We’ll be expecting the EAP features to come a’ rollin’ out now.

Version 9 any day now…

Another benefit of Super Cruise is that it’s “Big Brother approved”.

I hope Tesla adopts the system that Cadillac is using to monitor driver attentiveness. I think it’s pretty clear that Cadillac’s approach is better, both in terms of safety and not annoying the driver with constant “nanny” tests.

But overall, from all the reports I’ve read, I’d say it’s pretty clear that Cadillac Super Cruise functions better and smoother than Tesla Autopilot+AutoSteer within the limits of its capabilities. But as the article quite correctly points out, the capabilities of Super Cruise are far less than AutoSteer’s.

It seems strange to me to try to conclude that either is “better”. That’s like claiming either a Cadillac luxury sedan or a 4×4 pickup is “better”. It all depends on what you want to use it for. The Cadillac sedan will certainly give you a more comfortable ride so long as you stick to well-maintained paved roads, but the 4×4 pickup will take you places that you’d never dare take the Cadillac!

Basically one works and one will kill you. The problem with AP and for that matter all autonomous cars which are not a Level 4 or Level 5 is that you can’t manage the transition from autonomy to the human driver. Waymo figured this out as soon as they allowed employees to use the cars. They napped, looked at maps, fiddled with their phones, put on makeup, and generally “lost contextual awareness”. That made it impossible for the driver to take control of the car in dangerous situation. So no driver beats a distracted driver.

SuperCruise works because it can only be activated in limited situations and makes an effort to ensure the driver stays engaged. Going forward autonomous vehicles will use the same limited situations but dispense with the driver rather than trying to keep him/her engaged.

Neither system will ever be FSD (level 4 or 5) by doing incremental updates. This is where Tesla is losing the race. They are on a dead-end path.

Cruise Automation, Google, and Uber (and maybe Apple) will get there first. And now Honda is going with Cruise as well.

This is one area where Lutz is close to right. Tesla engineers have no special sauce. They just have a risk-taker for a boss.

If supercruise is so good,why isn’t it in the Bolt and other GM cars?

It is expensive. that is why it is only offered in one of the most advanced sedan GM has.

Bolt is being used to test the even more advanced system with Cruise Automation.

GM is planning to offer improved version of it across its lineup by 2020.