A bit of caution for autonomous vehicles

There has been lots of talk about autonomous vehicles. Driverless cars if you prefer that term. Even the general public is interested in the subject.

But I wonder if we are willing autonomous trucks into existence because of pressure by regulators to reduce accident rates, drivers less interested in driving cars, let alone trucks, companies wanting to make money off the sale of the systems needed to make autonomous vehicles a reality and the vested interest of researchers and NGOs.

I think there will be a painful learning curve on our way to fully autonomous vehicles especially if we push too far too fast. A key point to remember is that in an autonomous vehicle a less engaged driver — much like an airline pilot — must always be ready to take control. Recall the airplane accident in San Francisco airport or Tesla’s recent autopilot incident in Florida.

There are other factors to consider too, like the fact that there is no clear definition of what “minimum object detection” should be. Can the sensors on an autonomous vehicle spot a car in the roadway in any orientation? Can it spot an armadillo? A deer? A moose not moving? A pothole full of water or ice? Differentiate metal plates in the roadway? Confusing paint lines as occurs in construction zones where they paint black over the old dashed lines and add temporary new white lines? Deal with a sudden flat steer tire? 

I wonder if a driver crossing Kansas will get lulled into being distracted after an hour of autonomous driving so he will not be ready for that unexpected emergency which he is there to deal with. For example, a large tumbleweed blows across his path. Does the radar even pick it up? If it does, does it incorrectly slam on the brakes so the car following has to react? Does it swerve to avoid, again not expected by the tailgating car causing that driver to go off the road?

I am not anti-autonomous trucks and I do know that predictable incidents will be correctly handled by the autonomy and will give back control to the driver in a reasonable amount of time to handle the situation that the autonomy decides it can’t handle. This likely will work for nearly all driving situations. For the remaining, though, the vehicle software will fail — either unable to determine that it is at risk, or handing over control too late to be of value. 

Finally, though, there continues to be huge value in continuing the industry’s efforts on automating trucks.  We have integrated and advanced cruise control, refined automated manual transmissions and now there is Two-Truck Platooning, where two drivers steer and maintain control while the truck’s save about 4% in real-world fuel savings.  We studied Platooning and reported its status, benefits and challenges here.  Driverless, who knows when, but automating with drivers, we are ready now.

Hide comments


  • Allowed HTML tags: <em> <strong> <blockquote> <br> <p>

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.