By on May 22, 2019

tesla model 3

In this writer’s opinion, one of the greatest things to happen to high-speed motoring is the blind spot monitoring system. Try as we might to religiously check our mirrors and peer over our shoulders before each lane change, there’ll always be that time we half-ass it, just as an unseen car creeps up in the shadow of our B- or C-pillar. BSM can be a savior.

However, handing over the entire lane-change process to a combination of software and sensors, at least in Tesla vehicles, is far, far worse than doing it yourself, Consumer Reports claims. After giving the latest update to Tesla’s “Navigate on Autopilot” feature a shakedown cruise on the highways of Connecticut, the consumer advocacy group handed the system a failing grade.

The automaker rolled out an update to its Navigate on Autopilot feature in early April, claiming the system, which debuted in October 2018 and allowed Tesla vehicles to change lanes after a turn-signal prompt from the driver, to dispense with the bag of flesh and bones behind the wheel entirely. Only buyers who paid extra for Enhanced Autopilot or the misleading Full Self-Driving Capability gain the feature.

With the update, drivers can set the system to automatically change lanes when a destination point has been entered into the car’s navigation system. It can also be set to turn on automatically whenever a driver inputs a destination, rather than ask for permission each time. Hitting the turn signal stalk will defeat the looming lane change.

While drivers can choose to be notified of an impending lane change via audio, visual signals, or a wheel shaker, this would, in the minds of many tech-hungry drivers, defeat the purpose of having such a system. If you’re already checking your blind spots to ensure the computer isn’t making a mistake, why not just signal and move the wheel while you’re at it?

Out on the highway, Consumer Reports’ Model 3 test car reportedly behaved like BMW 3 Series, cutting off cars and passing on the right.

“As a result, the driver often had to prevent the system from making poor decisions,” CR stated. Once again, if you have to babysit your advanced driver-assist feature, what good is it to the driver?

“The system’s role should be to help the driver, but the way this technology is deployed, it’s the other way around,” said Jake Fisher, Consumer Reports’ senior director of auto testing. “It’s incredibly nearsighted. It doesn’t appear to react to brake lights or turn signals, it can’t anticipate what other drivers will do, and as a result, you constantly have to be one step ahead of it.”

The update made the test car no less erratic than when it was tested with the original Navigate in November, CR said. Even then, “it lagged behind a human driver’s abilities in more complex driving scenarios despite Tesla’s claim that it would make driving ‘more relaxing, enjoyable and fun.'”

Recall, too, that Tesla claims it will have fully self-driving cars by the end of the year.

In rolling out the April update, the automaker said drivers have already traveled “more than 66 million miles using the feature, and more than 9 million suggested lane changes have been successfully executed with the feature in use.” It isn’t known in how many instances a vehicle approaching from the rear had to brake hard to avoid hitting the slower Tesla that just appeared in its lane.

Consumer Reports’ roast of the company’s latest tech update will surely lead to another round of the on-again, off-again Tesla-CR boxing match.

[Images: Tesla]

Get the latest TTAC e-Newsletter!

Recommended

27 Comments on “Consumer Reports Slams Tesla’s Navigate on Autopilot Update, Calls System ‘Far Less Competent’ Than a Human Driver...”


  • avatar
    APaGttH

    It’s OK. Clearly, this is anti-Tesla bias, there is an OTAU coming that will update this version, and when the flux capacitor is added to the Tesla Model 3 as a $7500 option, the car will be able to fly.

    Roads? Where we are going we don’t need…roads.

  • avatar
    R Henry

    “it can’t anticipate what other drivers will do”

    Herein CR identifies the real issue. I believe AVs can only work in a closed network environment with other AVs.

    • 0 avatar
      mcs

      “AVs can only work in a closed network environment with other AV”

      I think the current technology would have trouble there too. Animals, weather incidents, road debris etc. You can’t really have an autonomous vehicle closed network. That being said, eventually, we will have AVs that are just fine in the real world. It just might be a ways off. It’s not going to be overnight.

      Personally, I’m okay with the current technology for stop and go traffic and empty limited access highways away from on and off ramps.

      • 0 avatar
        R Henry

        I am 52, and don’t believe I will live long enough to see AVs and human piloted vehicles on the same roads.

        • 0 avatar
          mcs

          @R Henry

          Hopefully, you’ll live well past that date. Should be within 20 years.

          • 0 avatar
            ToolGuy

            mcs,

            (This comment is intended to relate to your ‘Boston/paintball’ comments, but is placed too high on the page.)

            You mention infrastructure/situational concerns, which are very valid.

            After those are solved, we are likely to encounter additional challenges due to driver behavior. I predict that algorithms developed in California and Arizona will meet serious challenges when implemented in Ohio, Alabama and Florida.

            (My son mocks me for noticing and paying attention to state license plates, and even which county in my home state. I say it keeps me alive – lol.)

        • 0 avatar
          Kruser

          Just fly to Phoenix, rent a car and go to Chandler a short drive away. I imagine if you drive around a bit, you’re likely to see a Waymo vehicle hauling people around. Their pilot just went from 500 customers to 1000.

          • 0 avatar
            mcs

            @kruser:

            Just fly to Phoenix, rent a car and go to Chandler

            Just fly to Boston and head north on 1a, to route 60, then to route 1. You will not see a Waymo vehicle hauling people around. Bell Circle and some of the other traffic circles there will blow that piece of craps little electronic mind. Bell circle is a mutant intersection that is pure evil and the next rotary after that doesn’t follow any rules. It’s a free-for-all.

            How about after a New England Patriots game when cops using hand signals direct some traffic to the lanes left of the center of the road and it has to handle driving on the wrong side of the road for several miles?

            Some of us in the industry are designing AV systems for the real world. Designing and testing in AZ is like a defense contractor designing and testing a main battle tank against paintball players. If it survives a direct hit from a paintball, an artillery shell should be no problem, right?

      • 0 avatar
        285exp

        “Personally, I’m okay with the current technology for stop and go traffic and empty limited access highways away from on and off ramps.”

        The problem is that we know that a lot of people will use the technology like it’s level 4 when it’s actually level 2.

  • avatar
    theflyersfan

    I’ve written it before and I’ll write it again – there is no way in bloody hell that any amount of computing power, lasers, cameras, and blind luck can compensate for the sheer lunacy that happens at rush hour down I-75 in major construction zones (three of them in the greater Cincinnati area alone, plus more in Northern Kentucky) with the narrow lanes, no reflectors, bad weather causing all kinds of glare, semis weaving around, cars moving 55 next to cars moving 30 mph, and heading southbound on the northbound side of the freeway. The human brain with all of the computing power behind it, combined with the years of experience that many drivers have, has a tough enough time handing the stress of that area.

    There are too many drivers, too many cars, too much construction, too many variables, bad weather, and an outdated freeway system for any of this to be a realistic goal. I understand the whole “getting it in bits and pieces” like the traffic jam assist, and some forms of adaptive cruise control. Those work only with an attentive driver ready to pull the plug on the computer control anyway. But to turn things over to a HAL 9000 with the kind of driving environment we currently have…no way. I still trust my brain and my skills to make the best decisions on the highway in all conditions and not trust some system that can crash, be hacked, be out of date, or just have a bad day.

    • 0 avatar
      redapple

      GM and Ford want to save a lot of money? STOP autonomous driving research. Let someone else do the heaviy lifting. Once it is operational – if it ever gets there- then copy, borrow , buy a useful de bugged system.

      It s a money pit now and will be such for a long time.

      Also, do the same to maven and her clones and BEVs. Save the cash until others figure it out.

      • 0 avatar
        APaGttH

        A full 1/3 of GMs market cap value is based upon their research on autonomous vehicles. GM has HUGE investments in this space, and CR stated that Super Cruise was superior to Tesla’a Autopilot but ya, GM is saving a lot of money by not doing any research in this space.

        /eyeroll

      • 0 avatar
        civicjohn

        @redapple, absolutely. Let the market waste the money and bring the technology to the market. Why F or GM feel the need to make these investments is stupefying. I suppose it is their way to manipulate their stock price.

        IF it ever gets figured out, those who do will be ready to sell their wares to all takers. Maven, indeed.

  • avatar
    EBFlex

    Typical Tesla. Release a half baked and dangerous piece of software and let the general public be the beta testers. Makes amazing sense.

    This company needs to go out of business. They are a cancer and Musk is a fraud.

  • avatar
    sirwired

    I find it hilarious that anybody at all believes Tesla’s proclamations that full Self-Driving is coming in just a few months.

    A system that keeps t-boning 18-wheelers, rear-ending fire trucks, and driving straight into retaining walls is a system that is *years* from being ready to drive itself.

    • 0 avatar
      TimK

      Here in Arizona multiple companies years ago promised that autonomous cars and semis would be plying the roads “real soon now”. Waymo was supposed to have over 10,000 driverless vans on the roads last year and some Chinese outfit was going to run dozens of LTL truck ‘bots down I-17 and I-10 every day.

      Hasn’t happened. Why? Because the crap software doesn’t work outside of a controlled lab environment.

  • avatar
    Master Baiter

    An “auto-pilot” system that you have to constantly babysit is more stressful and unsafe than simply driving yourself, assuming a relatively competent human driver.

    I test drove a BMW X7 the other day with lane keeping assist. It was a total PITA, as every time I went to change lanes, the wheel would fight against me. Yes, you’re supposed to signal first, but what if you have to take evasive maneuvers quickly and don’t have time to signal?

    Blind spot warning and adaptive cruise control is all the automation as I can tolerate.

  • avatar
    aja8888

    I could just see it now…..TV ad for new tech in cars circa 2025:

    “New for 2025 model year!

    The all new (brand name) SUV with “ALL DRIVER CONTROL”! Dump your buggy, unsafe AI controlled car for the Real DRIVING EXPERIENCE! No worries about crashing into the sides of trailers or running pedestrians down in crosswalks! Be the first in your neighborhood with modern the all human controlled SUV!

  • avatar
    ronald

    “Recall, too, that Tesla claims it will have fully self-driving cars by the end of the year.”

    Aha, but which year? 2017? 2019? I’ll take 2061 (not really)

    This is all so very silly. The amount of money being wasted on this is mind boggling.

  • avatar
    Yankee

    Blind spot monitoring would not be needed if people knew how to properly adjust their mirrors! Instead every car I get in as a mechanic has mirrors adjusted so the driver can just see the back of their own car. The back of your own car isn’t going anywhere! I learned this trick a long time ago when I was young and drove interesting cars with big C-pillars: lean over perfectly sideways so your head is just touching the side door glass glass. Now adjust your mirror so you can just see the back of your car. Lean the same amount to the right and adjust the same. Your mirrors will now be further out than you are comfortable with. Now go go for a drive in the right lane and let someone pass you. When they disappear from your rear view mirror, you will pick them up in your side mirror. When they leave your side mirror, you will be able to see them in your peripheral vision. Presto! 360-degree vision and no need to crane your neck anymore when in a merge lane. About 2 out of 10 people I show this to think it’s the greatest thing since sliced bread. The other 8 would prefer to keep looking at the back of their car and swiveling their head.

  • avatar
    SCE to AUX

    “if you have to babysit your advanced driver-assist feature, what good is it to the driver?”

    Congratulations, CR, you just figured out that Autopilot is a Level 2 autonomous system. As such, it doesn’t even need to work. Driver vigilance is *required* – not optional – on a Level 2 system.

  • avatar

    A friend who works in the money unlimited black world tells me that they can’t get this working either. Air travel is automated….the plane can take off and land by computer…but the space a plane travels is a lot simpler than an automotive trip, where a lot of sudden things can happen.

    I file this under the Popular Science magazine cover telling me the flying family car is coming 5 years from now……

  • avatar
    sirwired

    The tragic thing is, Level 2 automation, if *actaully* sold and operated like that, is a *wonderful* driving tool. I use the LKAS and ACC on my ’17 CR-V every time I’m on the highway.

    The adaptive cruise and lane-keeping functions handle all the short-distance chores that are boring and fatiguing, while my senses can pay more attention to the things the computer is incapable of, like spotting a distant 18-wheeler on an entrance ramp, a crappy driver weaving out of their lane, or a traffic jam coming up in a few hundred yards.

  • avatar
    stingray65

    If Elon announced it you can count on the fact that: 1) it will take at least twice as long to reach the market as he says, 2) it will work half as well (or less) than he announces, 3) it will be 20 to 50% more expensive than he predicts, and 4) it will lose money.

    • 0 avatar
      mcs

      If we’re talking full level 5:

      “1) it will take at least twice as long to reach the market as he says”

      I think that’s overly optimistic. I’d say 5 times longer.

      2) it will work half as well (or less) than he announces

      Yeah, that’s a safe bet. Maybe a bit optimistic.

      3) it will be 20 to 50% more expensive than he predicts,
      Probably. That’s a safe estimate.

      and 4) it will lose money.
      Probably.


Back to TopLeave a Reply

You must be logged in to post a comment.

Recent Comments

  • HotPotato: Doesn’t signal his lane changes, gets angry when car tries to make him drive like less of an a-hole?...
  • HotPotato: I always wanted a GVX. The idea of “the GTI of Yugos” is too delightful not to run with. And...
  • SuperCarEnthusiast: At least the Koreans are still selling sedans in the U.S. marketplace. And they are doing so with...
  • HotPotato: exactamente.
  • HotPotato: welcome to capitalism and its discontents.

New Car Research

Get a Free Dealer Quote

Staff

  • Contributors

  • Timothy Cain, Canada
  • Matthew Guy, Canada
  • Ronnie Schreiber, United States
  • Bozi Tatarevic, United States
  • Chris Tonn, United States
  • Corey Lewis, United States
  • Mark Baruth, United States
  • Moderators

  • Adam Tonge, United States
  • Corey Lewis, United States