
Accession Number : AD0658862
Title : OPTIMUM LINEAR INTERPOLATION OF SAMPLED FUNCTIONS,
Corporate Author : NORTHEASTERN UNIV BOSTON MASS COMMUNICATION THEORY GROUP
Personal Author(s) : Schetzen,Martin
Report Date : 01 APR 1967
Pagination or Media Count : 73
Abstract : A study of linear interpolation of sampled random processes is presented in this report. A mathematical model of a sampler and of an interpolator are first developed. These models are used to derive general expressions for the meansquare interpolation error. The specific examples of zero and firstorder interpolation are used to illustrate the expressions. The problem of optimum interpolation is then formulated using the criterion that the meansquare error be a minimum. Explicit expressions for the optimum causal linear filter for interpolation using corrupted samples and expressions for the resulting minimum meansquare error are obtained. These results are illustrated by some specific examples of practical importance. A simple upper bound of the meansquare error is derived and a generalization of the sampling theorem for random functions is obtained by use of this bound. Although the important case of periodic sampling is emphasized in this report, the extension to a periodic sampling is given. (Author)
Descriptors : (*INTERPOLATION, *SAMPLING), OPTIMIZATION, INFORMATION THEORY, MATHEMATICAL MODELS, STATISTICAL FUNCTIONS
Subject Categories : Operations Research
Cybernetics
Distribution Statement : APPROVED FOR PUBLIC RELEASE