Author Topic: 6 leg walker, neat!  (Read 7671 times)

0 Members and 1 Guest are viewing this topic.

Offline AdminTopic starter

  • Administrator
  • Supreme Robot
  • *****
  • Posts: 11,703
  • Helpful? 173
    • Society of Robots
6 leg walker, neat!
« on: October 28, 2006, 07:19:07 PM »
shows neat things walkers can do that you dont normally think about . . .

this can move its abdomen in any direction:
http://www.youtube.com/watch?v=a52Nz3Jvp_w

this has force sensors on its legs for terrain adaptation (scroll like half way)
http://www.youtube.com/watch?v=Ont07BS3bBE

Offline tuttut89

  • Beginner
  • *
  • Posts: 5
  • Helpful? 0
Re: 6 leg walker, neat!
« Reply #1 on: January 07, 2007, 01:53:42 PM »
That first robot is now number two in the coolest robots that I have ever seen! (First is that snake robot) It looks very agile and doesn't look jerky and mechanical at all- I am very impressed by how fluid the motion is.

Offline amakusa

  • Jr. Member
  • **
  • Posts: 23
  • Helpful? 0
Re: 6 leg walker, neat!
« Reply #2 on: September 30, 2007, 01:32:02 AM »
how do they get such fluidity in there bots!
If you prick me, do I not bleed? Yeah Battery acid.

Offline crossroads1946

  • Jr. Member
  • **
  • Posts: 21
  • Helpful? 0
Re: 6 leg walker, neat!
« Reply #3 on: October 01, 2007, 10:18:52 AM »
I would assume by simply reversing the kinematic equations so that the end effector is the joint connected to the body of the robot. They know that if they keep the movements within certain limits, the feet wont move. Very well done robot.

-Baker

Offline oelewapperke

  • Jr. Member
  • **
  • Posts: 22
  • Helpful? 0
Re: 6 leg walker, neat!
« Reply #4 on: October 02, 2007, 07:17:27 AM »
how do they get such fluidity in there bots!

Easy - you use control loops, designed to allow for continuous movement ... how to do this ? And you control the force the engine exerts, in order to control different variables.

You enforce constraints on the movement, on speed, on position, on acceleration and you adjust the engine speed according to what needs to be done. E.g. if speed is getting too high, reverse the polarity on the torque that you control.

(d/dt position = velocity, d/dt velocity = acceleration, d/dt acceleration = torque)

You change the torque immediately (ie just set it to some value), and you calculate the acceleration (voltage) sent to the engine like a1 = a0 + torque * (t1-t0)

Everything will move with a continuous acceleration, which is very, very smooth. To lower speed, set the torque to a negative (but NOT too negative) value. To speed up, set it to a positive value.

If you really want to impress people, do this type of control on the torque, ie one extra step. The advantage is that people interpret continuous torque as standing still. Ie you can move around in a room without triggering anyone's attention as long as you keep the torque of your movement continuous. This is very, very weird when you look at a robot arm.

You clearly see it is changing position, yet somehow you just can't believe that it's moving. You cannot tell when it stops moving (really, you just don't see it), you can't tell when it starts moving. It just fades into the background.

 


data_list