Generating multiple analog noise sources from a single linear feedback shift register with neural network applications

Abstract

A method is presented for generating multiple, arbitrarily shift, pseudorandom bit streams from a single linear feedback shift register (LFSR). Each bit stream is obtained by tapping the outputs of selected LFSR cells and feeding these tapped cell outputs through a set of exclusive-OR gates. This enables many neurons to share a single LFSR, resulting in an acceptable small overheads for VLSI implementation. This technique has been incorporated in a neural network chip that uses stochastic settling. Other possible applications include the testing of high-speed digital communication systems. >