A linear time-varying approach to the problem of signal reconstruction from a finite set of irregularly spaced noisy samples

The problem of signal reconstruction from finite noisy data is formulated in terms of designing a linear time-varying (LTV) system in the frequency domain. The estimator obtained by means of this approach is shown to coincide with a regularized version of a well known solution to the noise-free signal reconstruction problem. The proposed frequency-domain LTV system design formulation is seen to give an alternative tool for understanding the technique of regularization