1
$\begingroup$

I want to make a real-time end-effector position control simulation of a 6 DOF robotic arm. To do so, I should solve the inverse kinematic problem, but due to the complexity of calculations, it is impossible to do that in real-time. As I know, industrial robots like KUKA or ABB can solve the inverse kinematic problem in real-time. Can anyone help me with the algorithms and explain those solutions to me?

$\endgroup$
2
$\begingroup$

Closed-form inverse kinematic equations are used to determine joint angles based on the end effector pose. These equations are well established in the literature. For example, the classic robotics books by Craig or by Paul show the equations for typical 6dof industrial arms.

$\endgroup$
8
  • 1
    $\begingroup$ By the way, with regard to the “not possible to do in real time” part of your question: I programmed 6dof articulated arm inverse kinematics on an 80386 processor in the early 1990s, and closed the loop at over 100Hz. Today’s microprocessors can achieve near real-time loop rates much higher than that. $\endgroup$
    – SteveO
    Apr 5 at 23:55
  • $\begingroup$ Thanks for your answer. You suggest to use an analytical solution to solve the inverse kindmatic problem. what about the numerical solution? Why indutrial robots do not use a look up table and pre determined values for the inverse kindmatic? Because I think that would be faster. $\endgroup$
    – Ehsan_Amp
    Apr 6 at 2:46
  • $\begingroup$ I don’t why you would use a numerical solution when there are exact equations available. $\endgroup$
    – SteveO
    Apr 6 at 2:48
  • $\begingroup$ It depends how accurate you want to be. An alytical solutions are mainly fasters and for a complex structure It is highly recommended to use analytical solution $\endgroup$
    – Ehsan_Amp
    Apr 6 at 2:53
  • $\begingroup$ The lookup-tables for compensation values are combined with the analytical inverse kinematics solution. You have a large lookup table to increase the precision of the analytical solution $\endgroup$
    – 50k4
    Apr 6 at 15:15
0
$\begingroup$

3 motors from the base decide the xyz position of the end effector and the other 3 on the tip decide the rotation of it. This is a really simple model as Steve mentioned. I don't know how much speed do you need but it will easily reach 10kHz and over to numerically solve the inverse kinematics with a modern computer and modern lib such as ceres.

$\endgroup$

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.