Two simple models are provided showing the characteristics of basic iterative algorithms for the inversion of kinematics, namely the Jacobian transpose, its pseudo-inverse and the damped least-squares (DLS). The pro's and con's can be compared interactively for a serial two-links and a three-links chains. For the latter the gradient-projection method is also given to couple a secondary task exploiting the redundancy of the manipulator.
Ugo Pattacini (2020). Examples of Basic Iterative Algorithms for Inverse Kinematics. (https://www.mathworks.com/matlabcentral/fileexchange/29369-examples-of-basic-iterative-algorithms-for-inverse-kinematics), MATLAB Central File Exchange. Retrieved .
Thank you, just downloaded and learn from your simulation.It's a good example!
@Pedro, thanks a lot for spotting this out.
I've provided the fix in the latest release.
Hello. I downloaded and run your files. At first glance I realized that the example of the three_link robot was producing some strange behaviors. Browsing over the equations I see that you have a mistake in the Jacobian. You are feeding -sin(theta1) and cos(theta1) to the third component of the matrix. The correct term should be -sin(theta1+theta2+theta3), and cos(theta1+theta2+theta3). After these changes the example of the three link arm works perfectly fine. Great examples by the way.
very good work for 2 links and 3 links (redundant robot)
The bug is really fixed herein.
Fixed bug spotted by Pedro Miranda on the Jacobian computation of the 3-links chain example.
Just fixed a typo in the description
changes in Description field