0
$\begingroup$

I am trying to find the inverse kinematics formulation of DLR/HIT II hand. Till now I had success in finding an analytical method described in the thesis of Mavrogiannis, Christoforos I. named Grasp Synthesis Algorithms for Multifingered Robot Hands, given in appendix B. My question is regarding the A.28 where the author calculates q3 but has mentioned previously in the text that q3 is equal to q2.

Note: q here denotes joint angles

$\endgroup$
5
  • 1
    $\begingroup$ I cannot read that reference, but setting the two distal joints of the finger to equal angles is one of the typical ways to emulate basic human finger motion. It does not always hold true (we can use that joint redundancy to our advantage - just watch a guitar player!), but often does. See, for example, this paper: cdn.intechopen.com/pdfs-wm/46559.pdf $\endgroup$
    – SteveO
    Jan 16 '16 at 20:17
  • $\begingroup$ Thanks for your comment SteveO. In the text, it is given that each of the 5 finger has 4 joints[ q0,q1,q2,q3 ] and 3 degrees of freedom. q0 associated with abduction-adduction, q1 and q2 for flexion and q3 being equal to q2. My question was regarding the IK formulation where the author has calculated q1, q2 and q3 leaving q0, which is the source of my confusion. $\endgroup$
    – Utkarsh51
    Jan 17 '16 at 7:59
  • $\begingroup$ You're welcome. It has been over twenty years since I solved some stable grasp problems with the redundant hand, so there may be better approaches now. I had to throw the equations into a Simplex algorithm to give the least-squares optimized solution to the redundancy. Not sure if that helps or not. $\endgroup$
    – SteveO
    Jan 18 '16 at 3:20
  • $\begingroup$ i am trying to solve the same problem. Hope you have been successful in finding the inverse kinematics. I have a question which i hope you can clarify. In the thesis document, what are the variable l1,l2 used in (A.5) for finding the inverse kinematics in appendix B. It has not been stated anywhere. Thank you in advance. $\endgroup$
    – user13617
    May 5 '16 at 15:43
  • $\begingroup$ l1 and l2 are link lengths. Their values are l1= 55mm and l2= 25 mm. Hope this helps. $\endgroup$
    – Utkarsh51
    May 7 '16 at 10:13
0
$\begingroup$

It looks like maybe the author ditched the zero-based numbering, but I agree that it is hard to tell.

Why not try to contact the author directly? His LinkedIn page shows he's currently a graduate research assistant at Cornell. I've contacted authors in the past and everyone I've gotten in touch with seemed pleased that I did so.

$\endgroup$
3
  • $\begingroup$ Thanks for the suggestion @Chuck, I would contact the author. $\endgroup$
    – Utkarsh51
    Jan 18 '16 at 8:23
  • $\begingroup$ I contacted the author and this was his reply "In the derivation of the inverse kinematics, I have probably made a mistake in my notation after equation A.4. So, in my calculations q_1 is actually DLR's q_0, q_2 is q_1, q_3 (coupled with q_4) is q_2. " $\endgroup$
    – Utkarsh51
    Jan 21 '16 at 16:32
  • $\begingroup$ @Utkarsh51 - Excellent! Glad you were able to get in touch, and the problem was what it seemed; just a numbering mistake. $\endgroup$
    – Chuck
    Jan 21 '16 at 18:50

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.