Signal reconstruction through minimization of a bandwidth metric

This paper describes a means of reconstructing signals through minimization of a bandwidth metric. The bandwidth metric relies on the specification of a weighting function. The ability to successfully reconstruct the missing samples of a signal is dependent on the functional form of the weighting function. Methods are developed whereby optimal weighting functions can be derived. Finally, the derivation is given for a lower bound on the error of a signal. The methods developed can be used for both interpolation and extrapolation. The reconstruction technique is applicable to a variety of signal models.