Interpolation of Discrete-Time Series

Abstract

Interpolation can be used to up-sample a sparsely recorded or simulated data set. Ideal interpolation models such as a Nyquist-bandlimited filter in the frequency domain and sinc function in the time domain are difficult to implement. Previous researchers proposed an interpolation scheme that interpolates data faithfully and preserves the spectral features of the original data. This approach has one possible limitation that the interpolated data may not pass through the original data points exactly. This technical note provides a modification of this scheme to ensure the interpolated data capture the original data points.