Accession Number : AD0740480

Title :   A Study of Weighted and Operator Modified LMSE Algorithms for Probabilistically Defined Inputs,

Corporate Author : BELL AEROSPACE CO BUFFALO N Y

Personal Author(s) : Gerhardt,Lester A. ; Walbesser,William J.

Report Date : 04 APR 1972

Pagination or Media Count : 5

Abstract : An algorithmic process is developed which minimizes the least mean squared error resulting from a set of input vectors applied to a system such as communications, control, or guidance. These systems are characterized by the need for continual updating in order to remain adaptive to their environment. The results reported in the paper have helped to establish a unifying thread among algorithmic techniques for using probabilistic inputs in such adaptive systems as pattern recognition, control systems optimization techniques, and sequential estimation theory. (Author)

Descriptors :   (*INFORMATION THEORY, STATISTICAL ANALYSIS), CONTROL SYSTEMS, PATTERN RECOGNITION, PROBABILITY, ALGORITHMS

Subject Categories : Cybernetics

Distribution Statement : APPROVED FOR PUBLIC RELEASE