alec-glisman
commented
2 years ago Remember, only the quaternion itself needs to have a unitary norm.
Check the definition of a unit norm again and make sure this is what I am doing in the Udwadia constraint algorithm
It seems that I may also just be able to renormalize the result directly.
I recently learned that unit quaternions are not closed under elementwise addition as normal Euclidian vectors are. This implies that time integration of the components is a bit trickier than I first thought.
I need to check how to properly get the unit quaternion "velocity" from "acceleration" and "position" from "velocity" in the second-order integration scheme.