Spectrum prediction in dynamic spectrum access systems

Eltom, H 2018, Spectrum prediction in dynamic spectrum access systems, Doctor of Philosophy (PhD), Engineering, RMIT University.

Document type: Thesis
Collection: Theses

Attached Files
Name Description MIMEType Size
Eltom.pdf Eltom.pdf Click to show the corresponding preview/stream application/pdf; Bytes
Title Spectrum prediction in dynamic spectrum access systems
Author(s) Eltom, H
Year 2018
Abstract Despite the remarkable foreseen advancements in maximizing network capacities, the in-expansible nature of radio spectrum exposed outdated spectrum management techniques as a core limitation. Fixed spectrum allocation inefficiency has generated a proliferation of dynamic spectrum access solutions to accommodate the growing demand for wireless, and mobile applications.

This research primarily focuses on spectrum occupancy prediction which equip dynamic users with the cognitive ability to identify and exploit instantaneous availability of spectrum opportunities. The first part of this research is devoted to identifying candidate occupancy prediction techniques suitable for SOP scenarios are extensively analysed, and a theoretical based model selection framework is consolidated. The performance of single user Bayesian/Markov based techniques both analytically and numerically. Understanding performance bounds of Bayesian/Markov prediction allows the development of efficient occupancy prediction models. The third and fourth parts of this research investigates cooperative decision and data-based occupancy prediction.

The expected cooperative prediction accuracy gain is addressed based on the single user prediction model. Specifically, the third contributions provide analytical approximations of single user, as well as cooperative hard fusion based spectrum prediction. Finally, the forth contribution shows soft fusion is superior and more robust compared to hard fusion cooperative prediction in terms of prediction accuracy. Throughout this research, case study analysis is provided to evaluate the performance of the proposed approaches. Analytical approaches and Monte-Carlo simulation are compared for the performance metric of interest. Remarkably, the case study analysis confirmed that the statistical approximation can predict the performance of local and hard fusion cooperative prediction accurately, capturing all the essential aspects of signal detection performance, temporal dependency of spectrum occupancy as well as the finite nature of the network.
Degree Doctor of Philosophy (PhD)
Institution RMIT University
School, Department or Centre Engineering
Subjects Networking and Communications
Wireless Communications
Signal Processing
Keyword(s) Spectrum Prediction
Dynamic Spectrum Access
Cooperative Prediction
Hard Fusion
Soft Fusion
Hidden Markov Model
Bayesian Prediction
Version Filter Type
Access Statistics: 52 Abstract Views, 4 File Downloads  -  Detailed Statistics
Created: Fri, 06 Jul 2018, 14:57:51 EST by Denise Paciocco
© 2014 RMIT Research Repository • Powered by Fez SoftwareContact us