AVS 2016: Table Set for Tesla Tussle

AVS 2016: Table Set for Tesla Tussle

As predicted, by me, the anti-Tesla and anti-autonomous vehicle forces are gathering in the wake of the recent fatal Tesla Model S crash in Florida.  The rising resistance arrives one week in advance of next week's Automated Vehicles Symposium in San Francisco - setting the stage for a spirited debate.

Consumer Reports has now joined the chorus or critics which already included Joan Claybrook, president emeritus of Public Citizen and former National Highway Traffic Safety Administration administrator; Clarence Ditlow, executive director of the Center for Auto Safety; Rosemary Shahan, president of Consumers for Auto Safety and Reliability; and John Simpson, privacy project director for Consumer Watchdog - all of whom collaborated on a letter calling for President Obama to stop his "administration's undue haste to get autonomous vehicle technology to the road" until enforceable safety standards are in place.

At the core of the debate is a struggle for control of or the presence of the steering wheel in the car.  The discussion has taken on mechanical, moral, ethical and philosophical dimensions and is not likely to be resolved by regulators.  But more on that in a moment.

Consumer Reports is urging Tesla to disable the automatic steering feature in its autopilot system.  The request is part of a broader critique of Tesla's autopilot system including its misleading name and the nature of its deployment in the market.  This is just the latest episode in CR's ongoing love-hate relationship with the Model S which it has praised with its highest safety ratings and criticized for reliability issues.

What is most interesting about the CR request is that CR completely takes for granted the fact that Tesla could readily delete this feature from its cars - a feat no other auto maker in the world is capable of duplicating today.  The assumption is that the cars will be more acceptable, and presumably safer, if the driver is forced to do all the steering.

CR wants auto steer in the Tesla to be reprogrammed to require drivers to keep their hands on the steering wheel - a requirement that continues to be hotly debated throughout the industry and by regulators around the world all the way to the United Nations.  It is no surprise that a host of steering wheel sensing systems are currently on their way to the market along with a wide range of driver monitoring systems.

CR wants Tesla to force its drivers to keep control - even though researchers keep telling us that the machines do a better job of driving than the humans.  Of course, Google wants to remove the steering wheel altogether.

CR is also asking Tesla to fully test safety systems before public deployment.  It is hard to imagine that Tesla failed to test its autopilot function before deploying it.  What CR has in mind is cars driving on tracks or other off-road settings.

Unfortunately, off-road testing in faux driving environments like Michigan's mCity are almost a complete waste of time.  The data gathered in such faked scenarios is virtually worthless for use in real driving situations.

It is sad, but true, that Google and Tesla have it right in testing on real roads in real driving situations.  There is simply no way to replicate or recreate one-of-a-kind driving events or novel situations including coping with intersections and pedestrians and weather.  The only way to tackle automated driving is to take it head on - pun intended.

Finally, CR wants Tesla to stop calling its driving assistance system "autopilot" since this is creating unrealistic expectations among consumers.  Tesla defends the name, indicating no inclination to change it or its deployment, by comparing it to the autopilot function in aircraft which is in use when conditions are appropriate.  This is an unfortunate comparison given the fact that airline pilots receive substantial amounts of training - something most drivers never do.

Some critics suggest that the capabilities of autopilot tempt drivers into misusing the technology with potentially fatal consequences.  The problem with this argument is that Tesla did not invent driving mischief.  Drivers have always done silly and dangerous things with cars.  The critics may as well ask Tesla to put an end to bad driving behavior.

The issue central to discussions at the Automated Vehicles Symposium 2016 is control of the steering wheel.  A cursory reading of the latest revision of United Nations Economic Commission for Europe "Uniform Provisions Concerning the Approval of Vehicles with Regard to Steering Equipment" (https://tinyurl.com/hqppz3y) makes clear the challenge at hand.

New technologies are introducing an increasing variety of sensor-based inputs to the process of controlling the steering of a car up to and including communications received from roadside equipment wirelessly communicating with a vehicle.  Driver assist systems within the vehicle are increasingly working to keep cars within their chosen lane and to avoid collisions.

At the same time, regulations require the driver to maintain contact with the steering wheel.  But in a scenario where steering is assisted by sensors and safety systems, a driver keeping his hands on the steering wheel is more likely to cause a crash - as in the case of the Cruise Automation driver who grabbed the steering wheel sending his vehicle into a parked car.

The regulations and the regulators are hopelessly lost in resolving this contradiction.  Our mind tells us to let the safety systems do their work, but our heart (and the regulations and safety advocates) tell us to keep our hands on the wheel.

If the scientists and regulators are correct - that the machines can drive better than the humans - then the regulations will have to be modified.  As I wrote last week, there is no turning back on the road to automated driving.  You should keep your eyes on the road, but it's time to take your hands OFF the wheel.  (Apologies to Jim Morrison of "The Doors.")

Mike McGurrin

Retired Transportation Consultant and Executive

8 年

Roger: In the long run, I think you are right that automated vehicles will prove far safer than human drivers. But the problem with SAE levels 2 and 3 is that they get us only partly there, and perhaps at the cost of increasing the occurrence of some types of accidents, since humans make terrible backup monitors. I'm not talking about the intentionally unsafe behaviors such as climbing into the passenger seat while autopilot is engaged, but rather any system that expects a human to be always vigilant, ready to take over in a fraction of a second, after many minutes (or even hours) of boring monitoring. And these human factors issues are well known AND have been demonstrated even on test tracks where the subjects knew they were being monitored. It may turn out that, as Lloyd George wrote, "There is no greater mistake than to try to leap an abyss in two jumps." It's too soon to know, but it's possible that widespread safe deployment of advanced systems may require a single leap from ACC / automated breaking / limited lane keeping to SAE level 4 operating on selected roadways, rather than a totally incremental path. I'm reminded of the "uncanny valley" in robotics and animation, but in this case, that valley can have fatal consequences. Will level 2 or 3 systems prevent enough crashes that the crashes they induce are tolerable to society?

Robert Clark

EMODE Outdoors: Electric Mobility Development for Outdoor Recreation Destinations and On-Site

8 年

Right with you Joe! Campus and more closed and "gated" communities to start makes the most sense to me. How about around and between theme parks in Orlando? Back to your mention of airports. Their perimeters being "soft targets" for terrorists has me believing there's a use case for shuttling from a secure perimeter offset from terminal grounds potentially.

Demetrius Thompson??

President Chief Executive Officer @ Global Mobile Alert Corporation. People, culture & safety Connected Mobility AI Solutions Technology #1 Make it About Safety First.

8 年

Humans are more intelligent then a driverless car. We should always count on our own thoughts and actions. Therefore autonomous vehicles should only be used as a means of assistance to the driver. Think safety first.

回复
Robert Clark

EMODE Outdoors: Electric Mobility Development for Outdoor Recreation Destinations and On-Site

8 年

I agree with most everything you say Roger. But whee I differ is the real problem that is the variety of road and driving conditions that can add or subtract layers of complexity - and the ultimate scenario where there is a mix of driven and driverless on the most complex roadways. Worse case scenario for me is the conditions right where I live, the Dallas/Ft. Worth Metroplex. We are experiencing unprecedented economic growth, and along with Austin are playing catch up with our roadways and infrastructure. Driving conditions couldn't be worse, and accidents abound. Will adding driverless vehicles into this mix (and mix masters) improve the scenario? Sorry, but I'm a bigger proponent of select roadways, campuses or corridor lanes for driverless to run in - to start, not just any and all roads (and uncontrolled conditions).

要查看或添加评论,请登录

Roger C. Lanctot的更多文章

  • Musk vs. China ... in Space

    Musk vs. China ... in Space

    Could it be, the only organizations that are capable of competing with CEO Elon Musk’s SpaceX are sovereign nations? It…

  • Tesla Fails; Robotaxis Prevail

    Tesla Fails; Robotaxis Prevail

    A line has been drawn in the sand of automotive safety. On one side of the line are camera-only or camera-centric…

    16 条评论
  • More Than One Man Can Fix

    More Than One Man Can Fix

    Brian Tegtmeyer has been coordinator of the National 911 Program for the National Highway Traffic Safety Administration…

    2 条评论
  • FCC’s C-V2X Approval: A Dream Deferred

    FCC’s C-V2X Approval: A Dream Deferred

    It wasn’t supposed to be this way. When automotive engineers set out to save lives, the skies should open, the heavens…

    5 条评论
  • Element Leads Mobility with Autofleet Acquisition

    Element Leads Mobility with Autofleet Acquisition

    An already dominant player in the global fleet business, Element added a massive new wrinkle and growth multiplier with…

  • Route 66 & Indra: A State of the Art Interstate

    Route 66 & Indra: A State of the Art Interstate

    Change happens slowly and then all at once. It sneaks up on you.

    1 条评论
  • Turnsignl: Something So Right

    Turnsignl: Something So Right

    Sometimes a new application arrives in the market that is so disruptive, so out of left field, that you just know auto…

    6 条评论
  • L.A. Auto Show: The Eve of Destruction

    L.A. Auto Show: The Eve of Destruction

    The automotive industry is like catnip for American politicians. The billions of dollars in economic activity.

    4 条评论
  • Travel Takes a $570 Toll

    Travel Takes a $570 Toll

    It’s that time of year when Americans look longingly toward the open road and get their recreational vehicles cleaned…

    3 条评论
  • iHeartMedia's Heartless Stance on Radio

    iHeartMedia's Heartless Stance on Radio

    I have noticed recently while tooling around in the back of Lyfts that most drivers keep the radio on - at varying…

    3 条评论

社区洞察

其他会员也浏览了