Date of Award
Spring 2017
Document Type
Thesis
Degree Name
Master of Science (MS)
Department
Electrical & Computer Engineering
Committee Director
Khan M. Iftekharuddin
Committee Member
Chunsheng Xin
Committee Member
Jiang Li
Abstract
In this work, machine learning tools, including recurrent neural networks (RNNs), matrix completion, and non-negative matrix factorization (NMF), are used for cognitive radio problems. Specifically addressed are a missing data problem and a blind signal separation problem. A specialized RNN called Cellular Simultaneous Recurrent Network (CSRN), typically used in image processing applications, has been modified. The CRSN performs well for spatial spectrum prediction of radio signals with missing data. An algorithm called soft-impute for matrix completion used together with an RNN performs well for missing data problems in the radio spectrum time-frequency domain. Estimating missing spectrum data can improve cognitive radio efficiency. An NMF method called tuning pruning is used for blind source separation of radio signals in simulation. An NMF optimization technique using a geometric constraint is proposed to limit the solution space of blind signal separation. Both NMF methods are promising in addressing a security problem known as spectrum sensing data falsification attack.
Rights
In Copyright. URI: http://rightsstatements.org/vocab/InC/1.0/ This Item is protected by copyright and/or related rights. You are free to use this Item in any way that is permitted by the copyright and related rights legislation that applies to your use. For other uses you need to obtain permission from the rights-holder(s).
DOI
10.25777/7xga-r535
ISBN
9780355096392
Recommended Citation
Glandon, Alexander M..
"Recurrent Neural Networks and Matrix Methods for Cognitive Radio Spectrum Prediction and Security"
(2017). Master of Science (MS), Thesis, Electrical & Computer Engineering, Old Dominion University, DOI: 10.25777/7xga-r535
https://digitalcommons.odu.edu/ece_etds/16