Non-coherent PN code acquisition in direct sequence spread spectrum systems using a neural network

Abstract

An artificial neural network is described which performs parallel matched filtering of a received direct sequence spread spectrum (DSSS) signal corrupted with noise with a locally generated but time offset version of the received sequence. The network provides the complete cross-correlation of the local sequence and the received signal. The network structure and design procedure are described. Its performance in additive white Gaussian noise (AWGN) is evaluated and shown to compare very well with the theoretical performance of matched filter receivers. The purpose of the DSSS receiver is to despread the received signal and to remove the information content from the despread signal.<>