Numerical analysis of non-uniform sampling problem (Q1590789)

The author gives an overview of recent developments in the problem of reconstructing a band-limited signal from nonuniform sampling from a numerical point of view. A successful reconstruction of a signal \(f\) from its samples \(f(t_j)\) requires a priori information about \(f\), otherwise the reconstruction problem is ill-posed. In many applications, the signal can be assumed to be band-limited. Under this restriction, the reconstruction problem is well-posed and leads to efficient numerical algorithms. Furthermore, a fast multilevel algorithm is presented that can reconstruct signals of unknown bandwidth from noisy nonuniformly spaced samples. Efficient regularization methods for ill-conditioned reconstruction problems are also discussed. Examples from spectroscopy and geophysics show the great importance of the proposed methods.