4 ms·

BugJuggler: plans for a 21 m tall robot that juggles cars

21 points by unwind 12y ago

abcd_f 12y ago
Not built yet.
- unwind 12y ago
  True, I edited the title to make that clearer and less click-baity. Thanks.
markild 12y ago
Though an interesting concept, from an engineering perspective, this sounds like a horribly bad idea. Also, the project page doesn't seem to clearly get across how incredibly early stage it is. Their 8ft small scale robot arm is still conceptual...
- jack-r-abbit 12y ago
  What a joke. Their "prototype" video even says "Robot arm imagery is computer generated." How is that even a prototype?
StavrosK 12y ago
That "haptic interface" is never going to work. How can you juggle without feeling your hands?
- ars 12y ago
  Why can't it give force feedback?
- tlb 12y ago
  It's extremely hard at human scale. I worked on building human-sized walking robots, initially with hand controls, and I couldn't learn the eye-hand coordination needed to make it balance. Juggling is similar to walking in terms of speed and complexity of feedback. However, the fact that it's so large means the time constants are longer. At 10x bigger than a human, each juggle should take sqrt(10) as long. So maybe it's possible to do with entirely visual feedback.
nodata 12y ago
Hmph. Unless it can roller-skate too, I'm not interested.
zardeh 12y ago
As a reasonably proficient juggler and a less than reasonably proficient roboticist: The only functional juggling robot at any scale is housed inside a case and doesn't use an arm, it can juggle at exactly one width, and doesn't use any form of complex reverse kinematics. There isn't any kind of "arm" on any scale that is currently capable of juggling (that I've seen anyway). Some get close, but the only videos are of slower forms where they get assistance from people or interact with people. That also entirely ignores the physics of the situation. To get a 1000kg car to move up the 8 or so meters from catch to head height, and then catch it again, the forces are enormous. You'd need to apply nearly 80K newtons of force over a second or so to slow the object down on the catch.
mikkom 12y ago
Sounds safe.
chrisBob 12y ago
We had a hydraulic accumulator in the machinery for making MDF at work, and it scared the crap out of me. Priming it involved filling the reservoir with roughly 10 standard high pressure nitrogen bottles, and then compressing that further. The basic idea is that it is an air spring that you use to store energy. I am ok with sitting in the stands while cars fly through the air, but I am not comfortable knowing that these three guys are going to make a big breakthrough storing that much energy, in a very unstable form and then put it near paying customers. I have no problem believing that the first stage of the project can work, but this is the sort of thing that won't scale that well.

21 points by unwind 12y ago

abcd_f 12y ago

Not built yet.

unwind 12y ago
True, I edited the title to make that clearer and less click-baity. Thanks.

markild 12y ago

Though an interesting concept, from an engineering perspective, this sounds like a horribly bad idea. Also, the project page doesn't seem to clearly get across how incredibly early stage it is. Their 8ft small scale robot arm is still conceptual...

jack-r-abbit 12y ago
What a joke. Their "prototype" video even says "Robot arm imagery is computer generated." How is that even a prototype?

StavrosK 12y ago

That "haptic interface" is never going to work. How can you juggle without feeling your hands?

ars 12y ago
Why can't it give force feedback?
tlb 12y ago
It's extremely hard at human scale. I worked on building human-sized walking robots, initially with hand controls, and I couldn't learn the eye-hand coordination needed to make it balance. Juggling is similar to walking in terms of speed and complexity of feedback. However, the fact that it's so large means the time constants are longer. At 10x bigger than a human, each juggle should take sqrt(10) as long. So maybe it's possible to do with entirely visual feedback.

nodata 12y ago

Hmph. Unless it can roller-skate too, I'm not interested.

zardeh 12y ago

As a reasonably proficient juggler and a less than reasonably proficient roboticist: The only functional juggling robot at any scale is housed inside a case and doesn't use an arm, it can juggle at exactly one width, and doesn't use any form of complex reverse kinematics. There isn't any kind of "arm" on any scale that is currently capable of juggling (that I've seen anyway). Some get close, but the only videos are of slower forms where they get assistance from people or interact with people. That also entirely ignores the physics of the situation. To get a 1000kg car to move up the 8 or so meters from catch to head height, and then catch it again, the forces are enormous. You'd need to apply nearly 80K newtons of force over a second or so to slow the object down on the catch.

mikkom 12y ago

Sounds safe.

chrisBob 12y ago

We had a hydraulic accumulator in the machinery for making MDF at work, and it scared the crap out of me. Priming it involved filling the reservoir with roughly 10 standard high pressure nitrogen bottles, and then compressing that further. The basic idea is that it is an air spring that you use to store energy. I am ok with sitting in the stands while cars fly through the air, but I am not comfortable knowing that these three guys are going to make a big breakthrough storing that much energy, in a very unstable form and then put it near paying customers. I have no problem believing that the first stage of the project can work, but this is the sort of thing that won't scale that well.