Accession Number : AD0701791

Title :   ESTIMATION OF RANDOM FILTER CHARACTERISTICS,

Corporate Author : HAWAII UNIV HONOLULU INFORMATION SCIENCES PROGRAM

Personal Author(s) : Gaarder,N. Thomas

Report Date : FEB 1970

Pagination or Media Count : 23

Abstract : The report considers the estimation of homogeneous random filter characteristics using simple estimators. First a linear operation on the sum of the random filter output and independent white noise is used to estimate a linear operation on the instantaneous transfer function of the filter. Choosing the transmitted signal x(t) and linear weighting g(t) to minimize the mean square error is shown to be equivalent to finding the cross ambiguity function that approximates a known function best in a weighted least-squares sense. A lower bound on the minimum error variance is also found. The results unify and extend the results of Sussman and DeLong and Hofstetter, as well as providing a readily interpretable form of the as yet unsolved minimization problem. The estimation of generalized second moment of the transfer function is also considered. A simple estimator structure is used and the resultant minimization problem is similar to the first problem. (Author)

Descriptors :   (*ELECTRIC FILTERS, INFORMATION THEORY), WHITE NOISE, STOCHASTIC PROCESSES, TRANSFER FUNCTIONS, LEAST SQUARES METHOD

Subject Categories : Cybernetics

Distribution Statement : APPROVED FOR PUBLIC RELEASE