For alternate optimization, I used SCA to deal with the non-convex parts of the constraint terms. The final result satisfies the constraint, but does not match the target at all (least squares problem)

6