rethink-forrest
commented
6 years ago Two quick notes: 1) The code pull-request for this is #33 2) It seems like I based the decaying logarithmic barrier function partial from these notes: http://s-mat-pcs.oulu.fi/~keba/Optimointi/OP_penalty_engl.pdf
The position IK solver works by using gradient descent combined with a barrier function. It would be great if we had some sort of technical reference to go over this math in detail.