Study Says Ford, Not Tesla, Leads The Autonomous Car Race

4 months ago by Eric Loveday 41

Ford Fusion testing out future autonomous tech in California today

According to a new study conducted by Navigant Research, Ford is #1 in the autonomous car race, followed by General Motors in the 2 spot.

This is an odd result, as we’ve yet to see autonomous Fords in volume on public roads, but apparently there’s a lot of behind-the-scenes work being done at the Dearborn-based automaker.

Ever more surprisingly, Tesla didn’t even make the top ten cut.

As Tech Crunch explains:

“The annual survey looked at 18 different companies, smeared across 10 different criteria, ranging from strategy, to core tech development, to manufacturing capability and staying power…”

Here’s the autonomous leaderboard graphic from Navigant:

Autonomous Leaderboard Graphic Via Navigant

And here’s the top ten list:

1. Ford
2. GM
3. Renault-Nissan Alliance
4. Daimler
5. Volkswagen Group
6. BMW
7. Waymo
7. Volvo/Autoliv/Zenuity
9. Delphi
10. Hyundai Motor Group

Source: Navigant via Tech Crunch

Tags: , , , ,

41 responses to "Study Says Ford, Not Tesla, Leads The Autonomous Car Race"

  1. Kdawg says:

    Other than Renault-Nissan, it appears the Asian car-makers at at the bottom of the pack. I wonder why they have not expressed much interest in autonomous cars?

    1. spinner says:

      A fair number of people who could afford it already have a driver and it’s a status symbol to be sitting in the back.

    2. Just_Chris says:

      I suspect this is a bit like university rankings where the researchers determine the result they want and then come up with metrics that make that result a reality. There is no way Ford and GM are massively ahead of Merc or tesla.

      1. Kdawg says:

        Here’s the criteria used. I didn’t read the whole report.. just the splash screen at Navigant.
        ———

        “This Navigant Research Leaderboard Report examines the strategy and execution of 18 leading companies developing automated driving systems. These players are rated on 10 criteria: vision; go-to market strategy; partners; production strategy; technology; sales, marketing, and distribution; product capability; product quality and reliability; product portfolio; and staying power. Using Navigant Research’s proprietary Leaderboard methodology, companies are profiled, rated, and ranked with the goal of providing an objective assessment of their relative strengths and weaknesses in the global market for automated driving systems.”

        1. Nix says:

          When it comes to the “partners” category, it looks like they might be handing out bonus points for companies using Mobileye. The one thing in common for the top 3 car makers on their list is that they use Mobileye.

          I’m not so sure about that.

        2. wavelet says:

          Every Navigant “report” I recall quoted in InsideEVs has been on the gamut between Silly and Bulls***. This is no different, and combines both.
          None of the top criteria are knowable for any specific company in any depth (“vision”? “strategy”?), and even to the extent they were known, they will obviously change a lot in the next few years. Stuff like “production strategy” is years early to have any data on. Ditto sales and marketing, since there is no system that comes close to full SAE Level-5, so it’ss impossible to know how it’ll develop market-wise. The regulatory landscape is also very complex.

        3. Jim Whitehead says:

          Amen brother. By their logic, the buggy whip manufacturers in the 19th century should have crushed Mercedes and Ford, because they had a long proven strategy and execution ability… with old technology, duh.

          Why are the study criteria and weights “proprietary?” If we saw the math, might we conclude the study was rigged to favor the insiders paying for it?

          When I was a grad student in an Ivy League school, working with professors on Sociology studies, I saw that when a prof. got big grant money from a source, his study conclusions pleased that source over 9 times in 10. Its an amazing coincidence! 🙂 .

      2. stimpy says:

        This reminds me of the Gartner study a few years ago funded by Microsoft that said Windows Phone would take over iOS in 2 years.

        Clearly laughable to anyone not being paid by the old guard.

        1. Kdawg says:

          “Consumers Reports rates the top 10 Reporting Agencies. Rated at #1 is Consumer Reports”. 😀

  2. Shawn Marshall says:

    livesaving technology.
    Insurance rates will plummet.
    Will it help or hurt intercity buses? Probably hurt in the long term although give an initial boost.
    It is a wonder now that rail traffic could not be automated down to the car level – especially where two tracks are available.

    1. Mr. M says:

      A lot of trains already drive autonomous. See metro trains in Japan (puppets in driver seats) or some highspeed railway trains in europe.

      I don’t geht what you mean to say. By automated down you mean it to be made more unreliable or complex or what???

  3. Roy_H says:

    Tesla is clearly the most public, and their approach of using customers as beta testers is risky in my opinion. But I think this large number of people/miles must be giving Tesla some sort of edge.

    I don’t know enough about the Tesla system but I suspect it is primarily camera based and it would make sense to me that these cameras be in pairs for binocular vision. This would give good depth perception and mimic human capability. I don’t see LIDAR as any advantage and I wonder if this ranking considers LIDAR to be necessary.

    1. Someone out there says:

      Number of miles isn’t everything. People driving up and down the same stretch of highway every day doesn’t add much usable data to the system. What is better, doing the same experiment 10,000 times or doing 100 different experiments?

      1. (⌐■_■) Trollnonymous says:

        There is an advantage to having many drivers pass the same direction all the time.

        Tesla records (by camera possibly with OCR) road conditions (read street work and closures) all the time. If many travel the same direction many know the direction is open. Then if a closure occurs or whatever, then the many that encounter the closure sends the data to the cloud (just a swag they use the cloud for it) and the rest of the clan in that area know there’s a closure or whatever.
        Getting data from many endpoints is better than getting data from a few to make a decision.

    2. David March says:

      You’re an idiot. If you think stereo cameras are better than lidar then something is seriously wrong with you. Might as well make them color cameras too right? C’mon kid, think before you type.

      1. (⌐■_■) Trollnonymous says:

        lidar has an inherent flaw.

        Exposed moving parts.

        1. Pushmi-Pullyu says:

          Waymo’s (formerly Google’s) self-driving cars have their spinning lidar scanner under a transparent dome. Not sure why you think that’s an issue.

          I have read one claim that the mechanism which spins the lidar scanner isn’t robust enough to stand up to years of use on a car. But even if that’s true, that merely means the bearing needs to be beefed up and/or shock absorbers need to be added. Nothing wrong with the basic tech, so far as I know.

          Speaking as a computer programmer, I find it odd that so many people seem to think that the goal of designing sensor systems for self-driving cars should be to get them to act like the human eyeball and the human brain’s visual processing center. On the contrary, I’d much rather rely on an active scanning system that isn’t blind at night, isn’t subject to optical illusions, and one which can see thru fog. In other words, one that uses real-time, active scanning using either infrared lidar or radar, to paint a 3D image of the car’s environment in real time.

          1. (⌐■_■) Trollnonymous says:

            IEEE was doing something to develop an algorithm to use both LIDAR and Camera that would accurately see better and read road signs.
            Not sure what happened to the effort though.

            Speaking of fog, I think it would be entertaining to see the Autobots navigate through thick fog.
            Or maybe boring.

    3. Dan says:

      Tesla’s system is not autonomous.

      It’s more analogous to visually augmented cruise control that is able to follow lanes. That difference has somehow been lost in how they have marketed the technology to consumers, sometimes to disastrous consequences even in highway driving. 90% of the difficulty in autonomous driving is in non-highway mixed use settings. That’s the actual problem to be solved.

    4. Pushmi-Pullyu says:

      “…these cameras be in pairs for binocular vision. This would give good depth perception and mimic human capability.”

      Well, it would if microprocessors and optical object recognition software had the superior visual image processing power of the human eye and human brain. Sadly, they are abysmally short of that very high standard.

      That’s why self-driving cars need active scanners such as lidar or radar. Those provide much more reliable data for the simplified sort of data processing which microprocessors and real-time 3D mapping and collision avoidance software are equipped to handle.

      1. Dan says:

        Exactly. With all the superior processing power that the human brain has, our vision capabilities can be easily tricked. It’s the main reason why magic shows and Escher exhibits exist. Orienting ourselves in a 3D world with a set of noisy 2D images is inherently messy. The reason why humans don’t have better sensors is evident. There is no reason for a machine to artificially restrict itself.

  4. Didier says:

    Presenting this study without explaining it is useless, it is written in English, so do you really need a French guy to (try to) explain it ?

    The article takes into account factors such as the capacity to sell *many* autonomous cars and therefore the quantity of cars actually sold (and Tesla sells a lot less cars than GM or Ford or Renault-Nissan…). It is the same for Waymo (Google) which is probably the best for autonomous cars but is quite badly ranked because Waymo is not selling any cars !

    The article also focuses on the *best* technology without taking into account the price, while the goal of Tesla is to make autonomous cars *now*. Just as an example the autonomous Nissan Leaf uses 4 LIDARS (and a dozen of cameras and more classical radars) so it “selfdrives” well but cannot go on sell because nobody would buy a 100 000 USD 30 kWh Leaf… Nissan’s autonomous Leaf also need high definition maps that do exist only for the road where this car has been tested. Etc.

    So whatever concludes this study, Tesla will be the first to deliver autonomous cars, even if of course the car will have to give back the control of the car when it cannot handle the situation.

    But I have no doubt that when LIDARs will be affordable Tesla will use some, in a future next generation of models S, X, 3, Y.

    In fact it is the same than for electric mobility, for most automakers it is “for later” because “research must still be done”, eg with fuel cells since batteries were supposed to be not good enough. For Tesla, the “all trip electric car” is for today, and it will soon be as autonomous as the budget of the buyer makes it possible.

    Furthermore it would be relevant to check if none of the authors of the study is working for a well noted automaker… Let’s say Ford since it is ranked at the top.

    To conclude, any ranking depends of the measure used… and for non scientific studies the measure used often depends of the authors (and of who he is working for).

    1. Roy_H says:

      “The article takes into account factors such as the capacity to sell *many* autonomous cars and therefore the quantity of cars actually sold”

      If that is true, then why isn’t Toyota near the top?

    2. Pushmi-Pullyu says:

      The purpose of most or all Navigant “studies”, if they even deserve such a label, is to promote whatever agenda those who paid for the “study” want to promote, usually stocks or futures; or to denigrate whatever the customer wants to denigrate.

      As Just Chris already said, it’s most likely that Navigant chose whatever criteria would yield the result they wanted. That’s why different Navigant “studies” vary so wildly in their predictions.

      The study I’d love to see is how much variance there is between what former Navigant “studies” predicted and how things actually turned out!

  5. Mil says:

    I’ve never heard of Navigant. Anyone know how much credibility they have? Bit weird how they don’t explain on what basis they came to that conclusion. To me, it seems no different than Mr Joe Bloggs coming up with a ranking.

    1. ffbj says:

      Navigant is trash. A paid for shill for big auto. Or whatever company pays them for their “studies”.

      1. Stimpacker says:

        They are a small outfit that specializes in EVs, like Lux Research. They like to publish sensational tidbits for marketing visibility written by green staff.

        1. ffbj says:

          No, I think they funded by big oil mainly and are consistently downplaying the impact evs will have on oil and gas: For instance crap like this:
          “Now, the main thrust of Navigant Research’s arguments resides in four points. One, conventional vehicle fuel efficiency is meant to increase 22% over the next decade resulting in significant oil displacement to belittle the oil displaced by EVs. Two, autonomous vehicles will also tend to increase fuel efficiency on the roads further contributing to displacing oil. Three, assuming oil prices stay in the $40-$80 range for the next 10 years, conventional hybrids are likely to win the energy cost equation over electric drive. And four, low oil prices may lead to reforms which will have a negative impact on EV sales.”

          Complete nonsense. They hire actual scientists to write their crap and then clothe it in scientific method, to try and give it some credibility.

          1. MikeM says:

            ” . . . conventional vehicle fuel efficiency is meant to increase 22% over the next decade”

            Oooh! 22% in only 10 years.

            Global warming fixed!
            Sign me up for a Hummer!

          2. ffbj says:

            Continuing in this vein:
            “According to Navigant, the majority of cars in 2017 will either employ a turbocharged or a supercharged engine, an electric motor, or a combination of an electric motor and a gasoline engine. “There is no single technology that will dominate fuel efficiency improvements over the forecast period through 2025,” says David Alexander, senior research analyst with Navigant Research. “The focus, instead, will be on incremental improvements in engines and transmissions, along with weight reduction in as many places as possible.”

            Navigant says that the most significant technology that has helped improve fuel efficiency in recent years has been stop-start technology, which reduces fuel consumption and cuts CO2 emissions by shutting a car’s engine down when the vehicle is not moving, and restarting it when the driver applies pressure to the gas pedal. In the report, Navigant predicts that of all new vehicles that will be sold in 2025, 58% will have stop-start technology.”

            I think we all know, or should, how worthless start/stop technology is, but just in case:
            https://www.youtube.com/watch?v=k159M8QhCIE
            (reviewer is a bit crude, and no friend of evs).

        2. Pushmi-Pullyu says:

          The “About” page at Navigant Research’s website claims “Navigant Research is a market research and consulting team that provides in-depth analysis of ‘global clean technology markets’.”

          So, their focus is on more than just EVs.

          But as has been said, their “studies” promote whatever agenda the customer pays them to promote.

          One should read Navigant “studies” the way one used to read the (sadly defunct) Weekly World News; for entertainment value, not for informational content.

      2. Vexar says:

        I’ll second that. I read the headline, then thought “I wonder if this is another Navigant Research study.”

        Eyup!

        It’s a marketing firm, far as I can tell.

  6. TL;TR. I think I don’t understand this study and its purpose. Is it to show how well traditional car manufacturers are prepared for the future? Are they trying to sell us on autonomous ICE cars? Do we really care?

  7. Loboc says:

    Tesla’s approach may be a dead end. Since they don’t have a fully autonomous car, they have no idea if the sensors/cameras being deployed now will allow that capability.

    They are climbing a mountain to reach the moon.

    1. Pushmi-Pullyu says:

      I don’t understand why Tesla keeps claiming that its sensor system will be adequate for fully autonomous cars. They need active scanning in all directions, not just the forward-facing fixed direction radars they’re using. Relying on cameras and optical object recognition for sensing to the sides and the rear ain’t gonna fly.

      We’ve already seen an unfortunate demonstration of the limitations relying on optical object recognition, in the one confirmed fatality in the (non-)driver of a Tesla Model S which, according to Tesla, confused the side of a semi trailer painted white with a “brightly lit sky”. That sort of confusion certainly wouldn’t happen with active scanning!

      1. Kdawg says:

        I think they are relying on AI to do better w/camera data, but it’s hard to emulate the human brain. Big data sets should help though.

      2. stan1 says:

        Humans largely rely solely on optical recognition. There no good objective reasons they cannot achieve better results using the same sensor data minus human processor distraction. That they use additional forward facing sensor technology certainly augments the possibility of better results.

  8. Bogdan says:

    I’ve only saw autonomous demonstrations from Tesla and Nissan so far.

    1. Mr. M says:

      Mercedes future truck 2020.

Leave a Reply