Signal reconstruction for non-equidistant finite length sample sets: a "KIS" approach

The reconstruction of a signal from a finite number of unequally spaced signal samples is considered. Depending on the assumed class of signals, several more or less optimal solutions are proposed in the literature. Mostly, these methods assume perfectly known sampling instants and perfect signal samples. The aim here is to determine the robustness of these methods to the measurement noise and to draw conclusions for their applicability in sampling scope timebase reconstruction.