It seems like there should be a set of best practices or something to reduce the iterations of the Newton-Raphson method to the inverse kinematics. Does anyone have any tips or tricks for choosing an initial guess for the joint variables when calculating the inverse kinematics? Or any papers that review/discuss/theorize this?

  • $\begingroup$ Like any problem I suppose the ICs are always annoying...but correct me if I'm wrong...don't you have an initial starting point to begin with...? Or do mean for subsequent calculations for other trajectories $\endgroup$ – morbo Aug 7 at 17:36
  • $\begingroup$ You have both an initial state of each joint, and also the desired transformation. The initial guess for the joint variables would be the starting joint angles closest to the desired transformation. As the Newton-Raphson is a method for nonlinear root-finding for the scalar x and θ. $\endgroup$ – Agent56289 Aug 7 at 17:52
  • $\begingroup$ Is the transformation in joint or cartesian? $\endgroup$ – tuskiomi Aug 9 at 20:49

Your Answer

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

Browse other questions tagged or ask your own question.