Classification of Correlation Signatures of Spread Spectrum Signals Using Neural Networks.

摘要

The major goals of this thesis were to determine if Artificial Neural Networks (ANNs) could be trained to classify the correlation signatures of two classes of spread spectrum signals and four classes of spread spectrum signals. Also, the possibility of training an ANN to classify features of the signatures other than signal class was investigated. Radial Basis Function Networks and Back-Propagation Networks were used for the classification problems. Correlation signatures of four types or classes were obtained from United States Army Harry Diamond Laboratories. The four types are as follows: direct sequence (DS), linearly-stepped frequency hopped (LSFH), randomly-driven frequency hopped (RDFH), and a hybrid of direct sequence and randomly-driven frequency hopped (HYB). These signatures were preprocessed and separated into various training and test data sets for presentation to neural networks. Radial Basis Function Networks and Back-Propagation Networks trained directly on two classes (DS and LSFH) and four classes (DS, LSFH, RDFH, and HYB) of correlation signatures. Classification accuracies ranged from 79% to 92% for the two class problem and from 70% to 76% for the four class problem. The Radial Basis Function Networks consistently produced classification accuracies from 5% to 10% higher than accuracies produced by the Back-Propagation Networks. The Radial Basis Function Networks produced this classification advantage in significantly less training time for all cases.