All About The Google Autonomous Vehicle Project

There have been many self-driving cars made with different levels of success, but probably the most well-known project is the Google car.  What you may not have heard of, though is the autonomous Google cart, or golf cart to be exact. The first video after the break explains the motivation behind the cart and the autonomous vehicle project.  As with another autonomous vehicle we’ve featured before, they didn’t forget to include an E-stop button (at 1:03)!

In the second video (also after the break) Google’s Sebastian Thrun and Chris Urmson get into more of the details of how Google’s more famous autonomous Prius vehicles work and their travels around different towns in California. A safety driver is still used at this point, but the sensor package includes a roof-mounted 64-beam laser sensor, wheel encoder, radars, and a GPS sensor. With Google’s vast resources as well as their work with Streetview and Google maps, it’ll be interesting to see what comes of this technology.  I, for one, welcome our new robotic overlords.

[youtube http://www.youtube.com/watch?v=rOWhu_aa9kM&w=470&h=315%5D

[youtube http://www.youtube.com/watch?v=YXylqtEQ0tk&w=470&h=315%5D

Via [ieee Spectrum]

32 thoughts on “All About The Google Autonomous Vehicle Project

  1. I want to see a google car handle traffic around Paris. We saw a motorcade going down oncoming traffic in the middle of the day. Also, the rotary around the arc du triomphe was an insane swarm of traffic.

    1. not everything is politics. if the person responsible for this wanted to do his presentation on a mac it’s his prerogative. i dont think google is gonna be like “no one here can use mac products!”

      as for the videos, i like it. i want one. (though i do notice that every car equipped with this system is essentially a google street car. their aggregating data about the world in tests is all good and well, but when it goes live to what degree will this data be aggregated and will that data then be sent to google HQ? personally i imagine that would be the case, each car giving up to date information about the world to google in order to maintain an up-to-date and safer driving system, the reverse of that of course is that information is being sent to google which is more than just the state of the world but also the position of pedestrians around the car, if all cars were like this the position of a given pedestrian could be tracked just about anywhere.

      unless of course the cars only send maps of the area sans pedestrians.

      its not that i don’t trust google, im just inherently cautious of giving anyone that amount of power.

  2. Man…people will do anything to avoid driving a Prius!

    My car has a standard transmission, no cruse control, and an engine behind the seats (where it belongs)…and that is how I want it.

    I LOVE DRIVING MY CAR…but I would be very happy if those who hate driving (and are thus inattentive and slow) would turn over control to a robot and GET OUT OF MY WAY!

    1. Then again driving slow is not an indicator of inattention or a dislike of driving, anymore than fast driving is an indicator of a driver that’s attentive. Just relax unless you are still jumping from one nut to the other in your great,great, great,great grand daddy x 2, it’s unlikely you will see the the day that driving by humans done away with. Even that may be a stretch.

  3. I saw this on TV last week.
    Steven Hawkins tech program on the BBC, probably on iplayer. Amazing seeing it on live roads and merging into traffic on a highway.
    Sorry cant find a link to the program ans it seems to have disappeared from my pvr.

    Gordon

    1. By the time if(when?) this comes about, surveillance cameras in in car sensors will determine what party caused the accident before the call is sent for robotic personal to respond to the scene.

    2. the person who is normally responsible, unless the car itself is responsible by some error in its guidance system, at which point the makers of the guidance system. but if a human was driving manually and crash it then responsibility is resolved normally (like it would be in an human + human collision) if its driving automatically then responsibility is resolved normally, but rather than the human being responsible, if that car is found responsible, the company is held responsible.

      and that is something any maker of automated cars needs to be prepared for. they make THEMSELVES responsible for what the car does (which is great for consumers because people hate responsibility)

      the question is: how does it handle car trouble.
      for instance, suppose the car failed to break because the car was out of brake fluid, or the tires were bald, or suppose the car is driving and gets a flat, or a catastrophic blowout. suppose someone put water in their radiator in winter and the engine explodes. suppose the car hits ice, can it tell the existence of ice? (certainly not something you will test in California) those are cases where the system will need tested, and its areas of grey responsibility (who is responsible if the car hits another car because the owner didnt replace the brake fluid, the car for not knowing it was low on brake fluid, or the owner for not replacing the brake fluid.)

      google still has some ways to go on this.

  4. There are 64 laser beams on this!?! It’s like these people are just inviting a “Terminator” scenario!!!

    Seriously though, it’s all fun and games until someone is blinded (or someones seeing eye dog is blinded).

  5. The Velodine LIDAR, the car’s primary sensor, is prohibitively expensive. Last I heard, it cost around 70k. So, no one will be “getting one” until Google replaces the Velodine with low cost sensors.

  6. Can’t they do this without the stupid spinning thing sticking up on the vehicle? A commercial vehicle with that knob on top will never sell. The knob presents itself as a prime target on a military vehicle.

  7. I’m curious , how this “robots” are (going to be) programmed in the case of an accident.

    I mean if there is detected that a crash is going to happen because for example of the place of some pedestrians and the velocity/direction of 2 other vehicles… how is going to act?

    From an ethically correct point of view it should take position that will result the less human injuries/deaths, from another ethically-“correct(?)” point of view it should try to pass the most damage-results to the one who seems responsible for the crash and not to the “innocents” (non-starters) of this accident.

    Or there could be a more egoistic reaction from a self/myowners-viewpoint or minimum-participation-viewpoint (for damage/cost/law/etc matters) which maybe make the accident worse for the others and near-zero-effects to itself.

    and to close that …
    this reaction should (not!?) be related to a price-factor of the vehicle [imagine a campaign “buy our car as it is MORE safe”, as most of the present campaign are]?
    is there going to be the same policy for every manufacturer [maybe hackable! ;) ] ?

      1. @dbear
        lol!
        but then why is he upset in the movie?
        we pass the ethical dilemmas to the machine and to the programmer of the machine! It wasn’t his disicion after all .. the best he could do was to report the bug

    1. If it calculates the direction to take for the most/least amount of damange, can you really call it an accident?

      Good thing murder can only occur when one human kills another human.

      1. @edonovan

        >If it calculates the direction to take
        >for the most/least amount of damage,
        >can you really call it an accident?

        Come on accidents happen all the time and not only by wrong human (robot?) decisions and mistakes.
        A machanical part can fail or a flat tire even to a non human driver.

        (And also in my example, maybe the car who causes the accident could be driven by a human, and the robot-driver is another car that detects the upcoming accident)

  8. It won’t matter. One or two good fatal accidents and the lawyers will put an end to all autonomous vehicles.
    Only way they will be on the road is if they can prove the autonomous cars are safer. After that it’ll cost so much for insurance to be able drive your own it won’t be worth it.

  9. I was in that conference audience, and it was a great presentation (and as a professional robotics researcher, I have to say that the technology is very impressive). A lot of people on here are saying “but what about…” and “how will it deal with…”. Trust me, they have this stuff well in hand. Whoever things that wires buried under the highway is the answer, give it up. The cost of doing that on new highways, let alone old ones (plus the massive potential for single-point failures) makes it a ridiculous suggestion.

Leave a Reply to kostas plataniotisCancel reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

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