A multivariate Powell-Sabin interpolant (Q937039)

The problem of constructing a \({C}^{1}\) piecewise quadratic interpolant to positional and gradient data defined at the vertices of a tessellation of \(n\)-simplices in \(\mathbb R^{n}\) is approached. The authors propose for the interpolation scheme to split the \(n\)-simplices in an appropriately tessellation such that subdivision points satisfy some geometrical constraints (as collinearity or coplanarity for \(\mathbb R^{3}\)). The particular case when the data points form a regular lattice is considered. For non-gridded data the authors propose to choose the subdivision point of every face of each simplex as its circumcenter (interior to the face) and prove that the constraints are satisfied. The constraints are constructed based on Bernstein-Bézier form for polynomials defined over simplices. Moreover, the authors prove that the proposed interpolant has full approximation power.