Very nice solution.
I immediately added them to the Collection of examples from discourse.threejs.org ==> CirclesOnCurve
I had tried quaternions for this problem as well: Struggling with paths - #10 by hofk
However, I didn’t get the position of the model right and solved it later with matrices. MovementOnCurve
I was interested to know if your variant would also be a solution. Therefore I modified the code a bit and added different radii to the ellipse. It can be seen that here, too, a distortion is taking place.
CirclesOnCurve - TEST
Conclusion: It is obviously only suitable for circles in this way.
What would have to be changed/added?