locked
Coordinate system in MRDS RRS feed

  • Question

  • Good time of the day to everyone!

    Currently I am working on creating of articulated Puma robot in MRDS. The model is ready and joints can move. And I have to program forward and inverse kinematics. You know that MRDS have the coordinate system like that: x (right), y (up), z (on us). But in all books Denavit-Hartenberg parameters are described in such a way that first joint have coordinate system like: x (on us), y (to the right), z (up).

    By making forward kinematics, I have to rotate coordinate system of the first joint, so that it should be equal to MRDS system. But I do not know how to rotate system in inverse kinematics, as they have already derived formulas and answer is in Axis space but not in Cartesian.

    My question is: Is it possible to change basic coordination system in MRDS? So that I could assign where x,y,z should be directed by myself?

    Thank you very much for answer.

    Monday, September 12, 2011 6:43 AM

Answers

  • No, you can't change the RDS coordinate system. Why can't you just figure out the coordinate transforms in each direction? Then you could transform all the coordinates, do your kinematics, and transform back again. Would that work?

    Trevor

     

    • Marked as answer by WTWIA Wednesday, September 14, 2011 6:48 AM
    Wednesday, September 14, 2011 6:41 AM

All replies

  • No, you can't change the RDS coordinate system. Why can't you just figure out the coordinate transforms in each direction? Then you could transform all the coordinates, do your kinematics, and transform back again. Would that work?

    Trevor

     

    • Marked as answer by WTWIA Wednesday, September 14, 2011 6:48 AM
    Wednesday, September 14, 2011 6:41 AM
  • No, you can't change the RDS coordinate system. Why can't you just figure out the coordinate transforms in each direction? Then you could transform all the coordinates, do your kinematics, and transform back again. Would that work?

    Trevor

     

    Hello Mr. Trevor,

    I did it sometime before. Exactly like you said it now. Perhaps decision will be interesting for someone:

    In forward kinematics one should rotate A0 matrix till the coordinate systems become the same.

    And in Inverse kinematics, as Mr. Trevor said, one should rotate point (x,y,z) from MRDS to DH coordinates, find solution and it will suit MRDS coordinate system.

    Wednesday, September 14, 2011 6:53 AM