Multivariate B-spline interpolation

I've run an FEM to gather a dataset consisting of an output in response to 3 predictor inputs. Is it feasibile to interpolate this dataset using B-splines? It seems the available functions are designed with a focus on univariate (and some bivariate) circumstances. So far I have been able to find the coefficients and specified knot locations, but have not been able to construct a model from which I can interpolate, which includes all 3 predictors.

Any help would be much appreciated,

Thanks!

Answer 1

[KSSV]

Read about interpn

[Thomas Rundle]

Thanks KSSV,

Unfortunately this has not really helped all that much. I can see how the interpn may assist in multivariate interpolation as a whole, however incorporating B-splines seems like a different challenge altogether... Maybe I'm not looking at this from the right angle?