No teen driving.

My dream is that if my kids can make it through college without a car, then by that point in the future, there will be no reason to learn how to drive.

Imagine skipping the teen driving years altogether. Year one is a death trap for all, and ADD teens are worse than drunk drivers in some studies. Imagine driving drunk for a year.

The robocars are on my mind and on NPR today as California just passed a law making them street legal. The opening video there is a pretty interesting place to start (stay with it through minute 2).

I was surprised to hear this part of the program:
“Ending Traffic Jams
Steve Jurvetson, a venture capitalist and self-driving car enthusiast, says safety is a huge benefit, but that’s just the beginning.

“Because we are going to go from about a billion cars on the road today to about 2 to 4 billion in the next 50 years, we can’t accommodate that in anything approaching the infrastructure we have in place,” Jurvetson says.

Picture global gridlock. If we don’t do something dramatic to enhance infrastructure or the way cars drive, Jurvetson says, the traffic jams will be unimaginable.

“But with autonomous cars, they can drive two to three times more densely,” he says. “You could, in fact today, remove all traffic jams from America if all cars went this way.”

The editorial chain lost the attribution for the forecasts — Bill Ford on global gridlock and Sebastian Thrun on the robo-solution.

25 responses to “Another Reason to Welcome Our Robocar Overlords”

  1. jurvetson Avatar
    jurvetson

    photos from my rides in the Google robo cars
    Google Robocar Racetrack Ride     … video too.

    In the meantime, some startups are developing software games to address the attention problem in teen drivers. In aged drivers, they found a 50% reduction in crash risk after a few hours of attention-honing games (cliincally designed and tested to promote neural plasticity within the sensory cortex).

    And it’s not just for the ADD extremes; 40-70% of teen accidents are explained by failures of attention or failures of visual search, and those crashes are often single vehicle collisions – the teen drives off the road and into an object. Like a telephone pole, which happened next to my house twice this year. They seem to be teen magnets.

    Reply
  2. WF&DT Avatar
    WF&DT

    It would be very nice if I didn’t have to handle another dead teenager for the remainder of my fire department career.

    Reply
  3. solerena Avatar
    solerena

    wow, 16 is not a good age to start driving for most kids, always felt this way…

    Reply
  4. msamaclean © Avatar
    msamaclean ©

    I’m all for the robocars! I hope I get to experience one in my driving life!
    (I actually saw my first one many moons ago @ the New York Worlds Fair….yes, old!…;-P…)

    Reply
  5. P^2 - Paul Avatar
    P^2 – Paul

    I present the contrarian view: The graph shows that drivers with more experience cause fewer crashes. So: Train our young ones to drive competently younger. They have the mental agility and reflexes to learn faster and better than, say, someone behind the wheel for the first time at age 21.

    Start them at 12 in a simulator game, and when they are 14 and big enough, in a full-size auto. By the time they are 18 and full of hormones and piss and vinegar they will have had a half decade of experience and will look like 21 year-olds on that graph.

    Reply
  6. linux-works Avatar
    linux-works

    ‘crashes per million miles driven’. and they list 16 year olds? am I to understand that some 16 year olds are driving a million miles or more??

    "can’t I stop yet, dad?"

    "no! keep going; there’s only about 800,000 miles left, son."

    😉

    Reply
  7. jurvetson Avatar
    jurvetson

    [http://www.flickr.com/photos/linux-works] – You know what they mean. Label the Y axis with x10^-6 and call it crashes per mile driven.

    [http://www.flickr.com/photos/imager] – Yes, there is an experience curve in every human activity, and the first year of driving is probably a bad one for any new driver. But I don’t see any reasonable simulators, or the movement to require a year of training in the Matrix. It might be easier to build a robo car than a realistic enough simulator. I say that because the robo car is basically done already. In my opinion, it is already a safer driving option than teens, and certainly ADD teens.

    P.S. photos from my rides in the Google robo carsGoogle Robocar Racetrack Ride … video too.

    In the meantime, some startups are developing software games to address the attention problem in teen drivers. In aged drivers, they found a 50% reduction in crash risk after a few hours of attention-honing games (cliincally designed and tested to promote neural plasticity within the sensory cortex).

    And it’s not just for the ADD extremes; 40-70% of teen accidents are explained by failures of attention or failures of visual search, and those crashes are often single vehicle collisions – the teen drives off the road and into an object. Like a telephone pole, which happened next to my house twice this year. They seem to be teen magnets.

    Reply
  8. pegleg000 Avatar
    pegleg000

    I don’t think it is a skill issue. I suspect raging hormones, peer pressure, and lack of judgement, that impairs the young driver. The judgement thing is more than can be learned by driving experience- it’s a more universal judgement. The same lack of judgement that can cause poor selection of friends, study habits, drinking, etc, leads to unsafe/high risk driving. Can lead to hard lessons… some call that experience. Hormones…… well, a necessary "evil"! 😉 Been there, done all that…. =8-o

    @msamaclean Yep, I remember the 1964 NY World’s Fair and the robocars, etc. Not old, just experienced. 😉

    Reply
  9. P^2 - Paul Avatar
    P^2 – Paul

    I enthusiastically look forward to robocars: not just because they are safer but also because driving is such a monumental time sink. Robocars promise to eliminate traffic jams, and reduce fuel consumption while simultaneously decreasing transit times. But more important to me is that I can get useful stuff done while riding in that box if I don’t have to pilot it.

    But what do we do about the "last mile problem"? Robocars will ultimately assume the majority of driving, but what happens in the fraction-of-a-percent of the time when the bag of meat needs to take over? You now have a human in control who may have minimal experience driving, but now needs to control the machine in a case where by definition it’s too complicated for the computer to handle. Do you give up? Hand control to a remote expert driving service? Implement some "cripple mode" computer-mediated control that limits speed and does crash-avoidence?

    It’s going to be interesting watching this evolve. Unfortunately, (tin foil hat mode on) it trivially enables remote monitoring and control, tracking, route selection favoring those who are "more important" or pay more, automatic blocking entry into areas or regions, automated tolls (probably a good thing), and generally greatly increased surveillance over the users. Gives me the willies thinking how this can be abused, it does. Google or not, great evil is possible here.

    Reply
  10. js.brain Avatar
    js.brain

    Can’t wait for them! When will the robocars be on the market in the US?

    Reply
  11. sbove Avatar
    sbove

    robo-mode in dense inner city traffic could be a great improvement…but when i’m on i50 in the middle of nevada heading for Telluride, it’ll be "robo-mode off, stealth mode on", cruise control set for 120+ 😉

    Reply
  12. Chucil Avatar
    Chucil

    Many happy faces for the hopefully reduced auto insurance premium. The elderly, the disabled, people with epilepsy, etc., will benefit greatly, too. Is China interested in this?

    Reply
  13. jurvetson Avatar
    jurvetson

    It can dodge objects — from boxes to deer to pedestrians — better than a human. It has a quicker reaction time and better sensory input. It can "see" things that we can’t see. When we were driving on Hwy 101, I marveled:

    "That front radar catches bounces off the ground," Jurvetson says. "We were driving behind an 18-wheeler, and we saw the vehicles in front of the 18-wheeler — vehicles we could not see with our eye — because the signal bounced off the pavement … at a glancing angle underneath the 18-wheeler. And so no human will ever have the amount of information that these cars have when they are driving." (NPR Feb 2012)

    Spinning Velodyne LIDAR on Roof and image it captures…

    Velodyne High-Def LIDAR Robocar

    And from April Fools (but the screen shot is real):
    Google racing video screen

    [http://www.flickr.com/photos/imager] — I agree on the police state risk. This really begs for an operational constitution for Google’s pledge to do no evil. And, of course, this is how the Robopocalypse begins…

    Reply
  14. seatonsnet Avatar
    seatonsnet

    Probably in some energy starved future… upon us soon if China motorizes, we will see a move back to public transport and large well connected urban areas filled with easily reachable goods and services. Till then, driver-less and electric cars will serve as a bridge from one mentality to the other.

    Reply
  15. casrose1 Avatar
    casrose1

    I have driven back and forth across USA and up and down many times . I always wanted to put my car on some track or hook it up to back of 18 wheeler and take me along ! It is the last hold out to caveman days !! I really liked Hyperloop Elon talked about in this interview a few weeks ago . I have studied gravity since 1980’s and energy fields etc so know what he is talking about is very doable and makes tons of sense ! People do not need individual cars !! People transporters utilizes what we have available. Land is gorgeous. No need for it to be taken up with all this hardware.

    My friend in Boston started City Water Taxi in 1990’s that utilizes Boston harbor to get to different part of Bean Town ! Much faster then land transport.. This has revitalized the harbor now and bringing the lost towns of east Boston, Chelsea, Charleston, south boston, revere etc into more of inclusive hub…

    http://www.businessweek.com/articles/2012-09-13/elon-musk-the-21...
    "On the assumption that people will be living on earth for some time, Musk is cooking up plans for something he calls the Hyperloop. He won’t share specifics but says it’s some sort of tube capable of taking someone from downtown San Francisco to Los Angeles in 30 minutes. He calls it a “fifth mode of transportation”—the previous four being train, plane, automobile, and boat. “What you want is something that never crashes, that’s at least twice as fast as a plane, that’s solar powered and that leaves right when you arrive, so there is no waiting for a specific departure time,” Musk says. His friends claim he’s had a Hyperloop technological breakthrough over the summer. “I’d like to talk to the governor and president about it,” Musk continues. “Because the $60 billion bullet train they’re proposing in California would be the slowest bullet train in the world at the highest cost per mile. They’re going for records in all the wrong ways.” The cost of the SF-LA Hyperloop would be in the $6 billion range, he says.

    Reply
  16. Jim Rees Avatar
    Jim Rees

    But we don’t yet know what the crash rate is for self-driving cars, do we?

    Reply
  17. ukweli Avatar
    ukweli

    [http://www.flickr.com/photos/imager] I strongly agree that the opportunities for abuse are just immense with these cars and systems. There is something deeply disturbing for me in any system that networks human physical locations. It’s primal for me, I can’t even identify exactly what about it bothers me.

    I see this taking a path more like a service than car ownership in the future. I imagine basically a "distributed bus" that picks you up when you send the service a signal from your phone. The maintenance on these vehicles will be critical and probably unattractive to own.

    Reply
  18. ukweli Avatar
    ukweli

    I wonder more about what the vehicle will do when it has a sudden mechanical failure than a road obstruction.

    I remember a day when my friend and his girlfriend experienced all 8 windows in their 2 VW Jettas falling into the doors because of a design flaw in the window supports. Imagine extrapolating an error like this into a self-driving car times one million.

    We may find out that we undervalue the decency and common sense of the human driver.

    Reply
  19. jurvetson Avatar
    jurvetson

    That sounds like a mechanical failure in the VW. A good system software layer might help with presumed failures in physical subcomponents (like a Google cluster or SpaceX engine cluster =)

    You are probably right on the distributed bus idea. It’s just a bit too radical to introduce when selling the initial appeal, but in the end, driving becomes an escort service. Google has modeled it, and for small EVs, the cost and environmental footprint is generally better than busses and trains. And the convenience would be door to door. With a few exceptions (in Japan for example), the train tends to ride mostly empty half the time in rush hour (most people witness the crush and not the counter-commute so more people experience trains as mostly full; so there is a bit of a sample bias there in popular perception).

    It might make sense in urban areas to start… an über-Uber if you will. Currently, most urban driving is spent looking for a parking space.

    Reply
  20. WF&DT Avatar
    WF&DT

    Given this discussion, it’s ironic that two days ago my engine company responded to and dealt with this fatal crash:

    Airpark Road

    Would a self-driving car have known that the road was wet and that it needed to slow down? The human driver did not, and a passenger in his vehicle paid the price.

    Reply
  21. ukweli Avatar
    ukweli

    If these systems achieve road presence dominance, it may have a large effect on certain types of crime. I imagine police being granted the ability to remotely slow, stop, lockdown, inspect by onboard camera, or reroute vehicles leaving the vicinity of a crime scene. I imagine "route" profiling. Drug trafficing would become very difficult, I imagine. Traffic violations would of course cease to exist.

    How will motorcycles fit into such a system? I imagine the man-machine interdependence of balancing a motorcycle would either massively complicate the control scheme, or remove the "man" element. The insurance costs for operating a manual motorcycle would probably become prohibitive pretty quickly.

    Reply
  22. Kaets Ebut Avatar
    Kaets Ebut

    > and ADD teens are worse than drunk drivers in some studies.

    Um…were you ADD? 😉

    (Sea-o-Rama!)

    Reply
  23. jurvetson Avatar
    jurvetson

    clearly! You see how that first year was just crazy.

    Reply
  24. Kaets Ebut Avatar
    Kaets Ebut

    FIRST year? Your crash graph is more like a step function!

    Reply
  25. jurvetson Avatar
    jurvetson

    Big steps in my first year…

    Let’s not forget the elder side of that crash curve… From today’s WSJ:

    "A government-funded study published this month found that playing Double Decision can slow and even reverse declines in brain function associated with aging, while playing crossword puzzles cannot."

    This follows "a multi-year, government-funded trial, known as ACTIVE, showed that participants followed for six years had a 50% lower rate of motor-vehicle accidents following cognitive training"

    Reply

Leave a Reply

Your email address will not be published. Required fields are marked *