• Doc Brown (unregistered) in reply to Mark
    Mark:
    Timing accuracy: If the robot doesn't close its grippers at the right moment, the block either falls out of open grippers or bounces off of closed grippers. Variation in network congestion is one of many variables that will significantly affect timing. Another two are the air pressure used to operate the gripper, and the temperature and viscosity of the oil that lubricates the pneumatic switch which actuates the grippers.

    Also, robot computers are generally not running a hard real-time operating system. This would be a necessity for any operation requiring sub-millisecond timing.

    While I agree that this story is bunk. The timing accuracy isn’t an issue with at least one current industrial communication protocol having the capability of deterministic 250 micro-second cycle times. And your other concers about timing lead me to believe you have not worked in a 24x7 production facility. Those concers are able to be taken care of with out too much trouble.

  • (cs)

    The stories are becoming more amateurish fiction over time, geez.

  • db (unregistered) in reply to zenstain

    Interesting little SF story, but somewhat ruined by the author not bothering to find out what a blast furnace is before mentioning it.

  • physicist (unregistered)

    What a pity that I'm too late for the party. But TRWTF for me is the fact that dozens of readers who probably had some science education can't come up with something better than guesses in >50 comments. Calculating a parabolic trajectory is not exactly rocket science (Yeah, ICBMs and the like - but you know what i mean). So here goes:

    Tossing an object at a 45° angle to a distance of 100' requires a take-off speed of about 27-28 mph. Depending on the shape of that metal block, air friction (assuming turbulent motion - which is pretty certain at that speed) will be between about 0.1% and 1% of gravity, causing a slight but computable slow-down. Top of parabola will be close to 25' (not 50', as someone else wrote) above start point. If there really were a "furnace" inside the factory hall, the ceiling would probably be high enough to not cause the projectile to bounce off of it. Flight duration is about 2.5 seconds. Assuming acceleration distance of 10' (dunno how long robot arms usually are), acceleration would be about 5 times gravity. Same for the catcher if it decelerates the projectile in 10'.

    As has already been pointed, the vision system would probably not be of much use for adjusting the catcher's arm in-flight, so we can safely assume it catches "blindly". That would require extremely high precision in time, speed (all 3 dimensions) and angular velocity. And if the catcher grabs the projectile from the sides (as opposed to catching it like a baseball), it would require extremely high grabbing force to prevent slipping. For a coefficient of friction between 0.5 and 2.0 and assuming the mentioned 5 G of deceleration, required force would be between 11 and 44 Newton per pound of projectile weight.

    Catching like a baseball without using a long deceleration distance would bruise (and maybe crush) projectile and robot arm badly, not to speak of the torque imparted on all bearings. Figuring out why/how much is left as an exercise for the reader ;)

    Nice fiction though. And would probably make an excellent problem for an undergraduate physics course.

  • (cs) in reply to physicist
    physicist:
    What a pity that I'm too late for the party. But TRWTF for me is the fact that dozens of readers who probably had some science education can't come up with something better than guesses in >50 comments. Calculating a parabolic trajectory is not exactly rocket science (Yeah, ICBMs and the like - but you know what i mean). So here goes:

    Tossing an object at a 45° angle to a distance of 100' requires a take-off speed of about 27-28 mph. Depending on the shape of that metal block, air friction (assuming turbulent motion - which is pretty certain at that speed) will be between about 0.1% and 1% of gravity, causing a slight but computable slow-down. Top of parabola will be close to 25' (not 50', as someone else wrote) above start point. If there really were a "furnace" inside the factory hall, the ceiling would probably be high enough to not cause the projectile to bounce off of it. Flight duration is about 2.5 seconds. Assuming acceleration distance of 10' (dunno how long robot arms usually are), acceleration would be about 5 times gravity. Same for the catcher if it decelerates the projectile in 10'.

    As has already been pointed, the vision system would probably not be of much use for adjusting the catcher's arm in-flight, so we can safely assume it catches "blindly". That would require extremely high precision in time, speed (all 3 dimensions) and angular velocity. And if the catcher grabs the projectile from the sides (as opposed to catching it like a baseball), it would require extremely high grabbing force to prevent slipping. For a coefficient of friction between 0.5 and 2.0 and assuming the mentioned 5 G of deceleration, required force would be between 11 and 44 Newton per pound of projectile weight.

    Catching like a baseball without using a long deceleration distance would bruise (and maybe crush) projectile and robot arm badly, not to speak of the torque imparted on all bearings. Figuring out why/how much is left as an exercise for the reader ;)

    Nice fiction though. And would probably make an excellent problem for an undergraduate physics course.

    Very good points. All of them.

  • Matz05 (unregistered)

    They probably just slid the parts across the floor. Would make a heck of a racket and give an observer the "THE ROBOTS ARE THROWING THINGS!!!!" reaction, but would be survivable enough for cast-iron parts and precise enough for the other robot to find and reach them.

  • sinni800 (unregistered)

    What if it wasn't feet but inches? ' can be confused with ".

    Just sayin'. 2.5 meters would be a lot easier to cross

  • Brian (unregistered)

    Replace "thrown" with "slid", and this is entirely plausible and pretty cool.

  • tlhonmey (unregistered)

    I don't know that this is as improbable as it sounds. If the joint that is imparting the velocity is capable of full 360 degree rotation, then it could, potentially, have a maximum velocity limited by friction and the strength of the bearings. At which point, the ability to throw an engine block that distance would become a question of how precisely the speed and release point could be controlled. Assuming the catching robot is similarly constructed, it could unwind the catch in a similar fashion, minimizing impact forces on both the arm and the block. As for using computer vision to fine-tune the catch: tracking moving objects in real-time is a common bit of demonstration code for OpenCV. Give it two (or, better yet, three) cameras from different angles and it could calculate fairly precisely what the actual trajectory of the block was, so it becomes just a matter of camera frame rate and having sufficient computational power to get the correction calculated and off to the catching robot before the block gets there. Some statistical algorithms to adjust the speed and release times based on previous throws and it should be able to fine-tune itself, and even correct for machine wear slowing things down.

    Still seems somewhat unlikely as a production-line story since the time required to write the code and calibrate it would probably be longer than just getting the conveyor repaired, and downtime is expensive... Perhaps the setting was adjusted slightly to further obfuscate the source and protect the guilty.

  • eric bloedow (unregistered)

    here's an amusing video i found with a robot throwing bowling balls 100 feet: https://www.youtube.com/watch?v=LAtdsDTt__s

  • Axel (unregistered)

    Hey! I KNOW this story is true!! For I AM the catching robot in this tale. Sure, my knuckles ached for a few days after that bitch shut me down, but otherwise, I'm fine. And I caught each and every one of those engine casings.

    But now, to post this comment, I must tap the box that says "I am not a robot."

    Oh, what bitter irony.

Leave a comment on “Caught”

Log In or post as a guest

Replying to comment #:

« Return to Article