Shop OBEX P1 Docs P2 Docs Learn Events
Inverse Kinematics Problem — Parallax Forums

Inverse Kinematics Problem

WaldoDTDWaldoDTD Posts: 142
edited 2006-02-07 03:36 in Robotics
hello,
Would someone please help explain to me how inverse kinematics works? At the last summer camp there was a robot arm which had a function that you could increment x,y and z functions and i am trying to do that with my crustcrawler 5DOF arm. Could someone give me some tips on how to learn this stuff? (I am taking Calculus AP now and i think that Jacobian matrices are a little over my head at this point). Thanks!-Signol01

Comments

  • GotenGoten Posts: 70
    edited 2006-01-26 03:14
    Hello Dude,

    ··· Here's some information... but's in spanish, if u don't understand·just ask
    ··· Direct Kinematic, Inverse Kinematic


    Best regards
    Goten turn.gif
  • Kevin WoodKevin Wood Posts: 1,266
    edited 2006-01-26 05:28
    Try this site:

    www.learnaboutrobots.com
  • WaldoDTDWaldoDTD Posts: 142
    edited 2006-02-04 17:02
    thanks, um... not really good on spanish (just finished spanish III in High School) if you don't have a spanish version i am sure i could get my teacher to help me translate. Also i have already been to that site and it only gives a rudimentary description of the problem not any solutions. Thanks!-Signol
  • Kevin WoodKevin Wood Posts: 1,266
    edited 2006-02-05 00:57
    Signol,

    From the above website, here is what I understand:

    Forward Kinematics:

    The forward kinematics problem is stated as follows: Given the angles at each of the robots joints, where is the robot's hand (Xhand, Yhand,
  • WaldoDTDWaldoDTD Posts: 142
    edited 2006-02-06 02:54
    This isn't making sense.
    I know how to do vector addition easy enough. This would give me an endpoint. Then when i try to do inverse kinematics (take the endpoint and work backwards to find the 3 joint angles) I have one equation with three unknowns which i have no clue how to solve.
    For instance i took a theoretical arm with 3 segments of 4, 3 and 1 units. The three joints would be set to the following angles, 50,130,130. This would then yeild for the x component of the point the equation 4cos50+3+cos50 =x. Then for the y component it would be 4sin50+0-sin50 (50 was gotten through the corresponding angles in two parallel lines and the fact that it mirrors the first joint). The end point was then found to be (6.21,2.3). Then when you work backwards I am stuck in the equation 4cosA+3cosB+cosC= 6.21 and 4sinA+3sinB+sinC=2.3. I am at a loss to where to go from here.-signol01
  • WaldoDTDWaldoDTD Posts: 142
    edited 2006-02-06 03:17
    Ok,
    Reread the site looks alot easier, have a few questions (dunno if you know just thought i would throw these out there) when it has the theta hand why is that parameter important in the end position of the hand? Can it be done without this parameter or is it essential? Also how would this parameter be determined? Also does the example go over a three jointed hand? If so would i have to modify it in any way if the third segment actually had a length? It looks as though it only uses a two segmented arm and i don't think that i could apply the law of cosines to three segments (i don't know its been a while since i went over that law in trig). Thanks!-Signol






    100 posts!!!!
  • WaldoDTDWaldoDTD Posts: 142
    edited 2006-02-07 03:36
    nevermind i figured out a way to eliminate one of the joints so that i can solve it regularly thanks!-signol01
Sign In or Register to comment.