Why driverless cars will screech to a halt

“Forget cures for cancer, climate change or world peace, the media has made it clear that self-driving cars will be the Next Great Step in civilization’s drive toward magnificence,” Steven Hill writes for Observer. “It’s time to hit the brakes for a reality check.”

“Despite how much Uber CEO Travis Kalanick likes to crow about our ‘driverless future,’ outside of The Jetsons this one is… not… happening… soon,” Hill writes. “Besides the remaining technological challenges, the liability and regulatory issues involved in letting a 3,000-pound death machine steer itself with no human at the controls are huge.”

“Think about it: Every driver makes hundreds of daily driving decisions that, strictly speaking, break driving laws (for example, crossing the yellow line to pull around a double-parked vehicle). It all works out fine because of something called ‘human judgment,'” Hill writes. “But what company is going to program its driverless cars to break the law? And what regulators will approve that product, knowing that it has been programmed to break the law?”

Much more in the full article here.

MacDailyNews Take: Thank you, Steven.

SEE ALSO:
Apple vs. Google in self-driving cars: To map or not to map? – March 6, 2015

49 Comments

  1. Will a driverless car be able to predict the tire wobbling and about to tear off from a big rig ahead? I know if I saw it I would get out of the way or even accelerate(above speed limit) to pass the situation. Not sure how a driverless car could see and realize what to do.

    1. You evidently have not seen the videos of the driverless cars dodging in response to objects suddenly appearing in the road ahead as well as real cars running a red light coming from a blind spot (in real life, not a “set up”. That is what infrared and radar like scanning does faster than humans can.

      By the way, I didn’t think this was possible and an anti-driverless advocate until a few weeks ago. Then I saw a couple of videos that were posted in links on MDN.

      1. You aren’t listening. Automated systems REACT. Humans PREDICT.

        That is why the best drivers will always be human, despite the fact that the horrid drivers training today on average turns out the least conscientious drivers I have ever seen.

        MOREOVER — it will take only a few incidents for lawyers to kill the profit motive for driverless cars.

        1. Come on, you aren’t thinking. SOME Humans predict based on input (called defensive driving). SOME! But Many don’t. Read my post on the first page. There are people like my wife that do not react defensively, as in predictive. REACTION auto detection would have prevented her wreck. She was in the right and defends that. But she could have avoided it. Automated systems would have prevented the wreck.

          Automated systems can “react” faster than most “thinking?” human drivers actually think/predict.

          Reaction driving: Do you know why the fighter jets are as sophisticated as they are today? The computers do far more flying than the pilots. You may not be old enough to remember experimental “flying wings” that were so delicate that humans could not safely and predictably fly them. Today’s stealth planes require computers to interpret the pilot’s input and push the plane to the limit (of that input) without putting the plane in aerodynamic danger. ERGO reactive computers are faster than predictive people in many cases. Today’s stealth planes (and even Harriers) are not naturally aerodynamic without predictive and reactive computers assisting/doing the flying.

          It is not an “either / or”, but: “both / and / either” as the need arises. A singular focus as to who, what or which – will soon be horse and buggy thinking.

          1. I guess my point is if the car has to react to the torn off tire is there enough reaction time to avoid the collision? If there isn’t enough then could it have been avoided? Prediction can go a long way and not sure driverless cars are capable of this. Please correct me if I am wrong.

  2. The trucking companies are very interested in this technology. Their biggest cost is drivers! Imagine the revolution this would present to the trucking industry if their loads can drive 24 hours straight and not have to stop for those silly human drivers to rest.

    Expect to see it happen in the next 10 years.

    1. Totally agree. Plus, self-driving trucks will be able to drive at speeds that optimize fuel-efficiency. They will not need to take human-required breaks to: rest, refresh, eat, sleep. And trucks won’t need to be tricked out with sleeping quarters for drivers. Their engines won’t need to idle for long periods (for whatever reasons they now do) while drivers do whatever they do at stops. And self-driving trucks will be able to travel at lower speeds than human drivers.

      Right now, human drivers travel at speeds to minimize total driving + running costs, basically: driver wages + fuel costs. But driverless trucks will be able to drive at speeds that minimize running costs: basically, fuel. Which should also minimize other routine costs for things like brakes, tires, oil, etc. It means driverless trucks will be able to drive at slower/safer travel speeds than seen with human drivers.

      Right now most long-haul truckers drive at the maximum speed limit. In the West, this means speeds up 80 mph, which clearly exceed the optimum fuel efficiency speed. Driverless trucks can roll at, say, 45 to 55 mph all week long without complaint.

      Driverless trucks can also be outfitted with smaller, more efficient, lower-powered engines. Even hybrid systems. Something like the truck equivalent of a Prius hybrid drive, with a small, efficient engine optimized for fuel-efficiency, coupled with an electric battery, and matched with an Electronic Continuously-Variable Transmission (which matches power demand and engine output).

      Driverless trucks will revolutionize trucking by sharply and radically reducing trucking cost-per-mile and reducing fuel consumption and CO2 production. It is coming. Same for buses. And cars.

      1. Ultra-efficient transcontinental transportation systems that required little or no driver cost have been here for over 150 years — they are called trains. They also make roadways much more efficient by taking the big slow vehicles off the asphalt and putting them onto safe, relatively traffic-free right-of-ways.

        And everyone with a brain can see that on a closed system, automation can be both safe and affordable. Trains have proven that for decades.

        But many decades ago, trucking companies and their customer corporations discovered that maintaining tracks cost money. They didn’t want to have to maintain so much infrastructure that was independent of the “socialist” interstate highway system. If roadways were subsidized by the public, then it would be cheaper to hire drivers to haul increasingly huge semis on public roads, far bigger than the tiny station wagons and pickup trucks that used to distribute goods from the rail station the last few miles to its final destination..

        That is why the road infrastructure in the USA is beat to crap and still corporations whine to every government to give them more relaxation of safety standards, allow Mexican trucks to cross the border with no inspections, for waivers on weight, an so forth. Meanwhile railways have one by one turned into bike paths.

        America would be dramatically improved if in the median of every divided highway there was a monorail freight system onto which all heavy rigs were obliged to use, getting them fully out of traffic. Automate the hell out of that.

        Do not pretend that automating cars that are designed for use on public thoroughfares that encounter everything from horses to pedestrians to suburban backstreet baseball games will ever be cost effective. The dream of widespread automated cars is just the latest hype from technologists who wrongly think that electronics can solve everything. Lawyers will have the last laugh on that.

  3. Expert Systems, which is what we sort of call ‘AI’ or ‘artificial intelligence’, are great at looking up data and responding to it. But driving involves expecting the unexpected; Watching for the loonies; Deferring to pedestrians who you’d swear have either a death wish or are looking for your collision insurance money.

    What’s missing is REAL AI. The ‘intelligent’, sort of, machines can certainly control the car’s functions. But they aren’t capable of full human perception, interpretation and judgement. That’s WAY too difficult for any so-called AI we have so far, except within that limited world of expert systems.

    I know Google uses their goofy, twirling camera self-driving cars as a Gee Whiz! marketing exploit. Would anyone actually find a reason to own one of those prototype thingies except for the Gee Whiz! factor? Not at this point in time.

  4. The problem is not driverless cars. It is driverless cars sharing roads with regular ones and their human drivers. The way to go about this is have zones devoted exclusively to driverless. those could be experimental at the begininng, moving on to busy downtowns and then on to less congested areas. An area with only driverless cars where ALL of them are programmed to follow the same directives will optimize everyone’s conmute time and safety with relatively few technological hurdles to overcome. It is sad to let go the fun of driving, but in an overpopulated world with ever scarcer resources it is one of the many sacrifices we’ll eventually have to make.

Reader Feedback

This site uses Akismet to reduce spam. Learn how your comment data is processed.