Learning-Based Spectrum Sensing and Access for Cognitive Radio Systems
A Master of Science thesis in Electrical Engineering by Menatalla Diaaeldin Shehabeldin entitled, "Learning-Based Spectrum Sensing and Access for Cognitive Radio Systems," submitted in June 2015. Thesis advisors are Dr. Mohamed El-Tarhuni and Dr. Khaled Assaleh. Soft and hard copy availabl...
Saved in:
| Main Author: | |
|---|---|
| Format: | doctoralThesis |
| Published: |
2015
|
| Subjects: | |
| Online Access: | http://hdl.handle.net/11073/7854 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1864513438427906048 |
|---|---|
| author | Shehabeldin, Menatalla Diaaeldin |
| author_facet | Shehabeldin, Menatalla Diaaeldin |
| author_role | author |
| dc.contributor.none.fl_str_mv | El-Tarhuni, Mohamed Assaleh, Khaled |
| dc.creator.none.fl_str_mv | Shehabeldin, Menatalla Diaaeldin |
| dc.date.none.fl_str_mv | 2015-07-02T09:39:32Z 2015-07-02T09:39:32Z 2015-06 |
| dc.format.none.fl_str_mv | application/pdf |
| dc.identifier.none.fl_str_mv | 35.232-2015.31 http://hdl.handle.net/11073/7854 |
| dc.language.none.fl_str_mv | en_US |
| dc.subject.none.fl_str_mv | Cognitive radio Spectrum management Dynamic spectrum access Spectrum sensing Hidden Markov model Polynomial classifier Nonlinear autoregressive with exogenous inputs model Cognitive radio networks Mathematical models |
| dc.title.none.fl_str_mv | Learning-Based Spectrum Sensing and Access for Cognitive Radio Systems |
| dc.type.none.fl_str_mv | info:eu-repo/semantics/publishedVersion info:eu-repo/semantics/doctoralThesis |
| description | A Master of Science thesis in Electrical Engineering by Menatalla Diaaeldin Shehabeldin entitled, "Learning-Based Spectrum Sensing and Access for Cognitive Radio Systems," submitted in June 2015. Thesis advisors are Dr. Mohamed El-Tarhuni and Dr. Khaled Assaleh. Soft and hard copy available. |
| format | doctoralThesis |
| id | aus_3554cbd8ed1840d7bbc943c1e203dc33 |
| identifier_str_mv | 35.232-2015.31 |
| language_invalid_str_mv | en_US |
| network_acronym_str | aus |
| network_name_str | aus |
| oai_identifier_str | oai:repository.aus.edu:11073/7854 |
| publishDate | 2015 |
| repository.mail.fl_str_mv | |
| repository.name.fl_str_mv | |
| repository_id_str | |
| spelling | Learning-Based Spectrum Sensing and Access for Cognitive Radio SystemsShehabeldin, Menatalla DiaaeldinCognitive radioSpectrum managementDynamic spectrum accessSpectrum sensingHidden Markov modelPolynomial classifierNonlinear autoregressive with exogenous inputs modelCognitive radio networksMathematical modelsA Master of Science thesis in Electrical Engineering by Menatalla Diaaeldin Shehabeldin entitled, "Learning-Based Spectrum Sensing and Access for Cognitive Radio Systems," submitted in June 2015. Thesis advisors are Dr. Mohamed El-Tarhuni and Dr. Khaled Assaleh. Soft and hard copy available.Spectrum management is one of the most important elements of the overall design of cognitive radio systems. Primary users (PUs), or license holders, should not be affected by the opportunistic use of the spectrum by the secondary users (SUs). Moreover, secondary users, or the non-license holders, should try to maximize their utilization of free channels for better spectrum efficiency. The decision whether to access a channel or not is crucial to both the primary and secondary users. In this thesis, an improved spectrum access algorithm is proposed for cognitive radio systems by modeling the primary user channel usage pattern as a Hidden Markov Model (HMM). The proposed algorithm maximizes the channel utilization without causing significant interference to the primary user by considering access based on the availability of the channel at the current time slot. The decision on the availability of the channel is investigated using three machine learning techniques, namely HMMs, polynomial classifiers and nonlinear autoregressive with exogenous inputs (NARX) models. Simulation results based on models from real spectrum measurements show that using the conventional HMMdecoding technique leads to high collision probabilities of around 25%. On the other hand, using polynomial classifiers for deciding the availability of the channel enhances the system performance significantly, with collision probabilities less than 1%, while maintaining high utilization probabilities. A thorough investigation of the effect of the order of the polynomial classifier shows that while lower orders reduce the computational complexity of the algorithm, higher orders are more robust to high levels of shadowing. Another approach to mitigate the effect of shadowing is using cooperative spectrum sensing, where multiple SUs send the sensing results to a fusion center, which makes a global decision about the availability of the channel. Results show that the decision based on the scores of the classifiers outperforms majority vote in terms of collision and utilization probabilities.College of EngineeringDepartment of Electrical EngineeringMaster of Science in Electrical Engineering (MSEE)El-Tarhuni, MohamedAssaleh, Khaled2015-07-02T09:39:32Z2015-07-02T09:39:32Z2015-06info:eu-repo/semantics/publishedVersioninfo:eu-repo/semantics/doctoralThesisapplication/pdf35.232-2015.31http://hdl.handle.net/11073/7854en_USoai:repository.aus.edu:11073/78542025-06-26T12:29:04Z |
| spellingShingle | Learning-Based Spectrum Sensing and Access for Cognitive Radio Systems Shehabeldin, Menatalla Diaaeldin Cognitive radio Spectrum management Dynamic spectrum access Spectrum sensing Hidden Markov model Polynomial classifier Nonlinear autoregressive with exogenous inputs model Cognitive radio networks Mathematical models |
| status_str | publishedVersion |
| title | Learning-Based Spectrum Sensing and Access for Cognitive Radio Systems |
| title_full | Learning-Based Spectrum Sensing and Access for Cognitive Radio Systems |
| title_fullStr | Learning-Based Spectrum Sensing and Access for Cognitive Radio Systems |
| title_full_unstemmed | Learning-Based Spectrum Sensing and Access for Cognitive Radio Systems |
| title_short | Learning-Based Spectrum Sensing and Access for Cognitive Radio Systems |
| title_sort | Learning-Based Spectrum Sensing and Access for Cognitive Radio Systems |
| topic | Cognitive radio Spectrum management Dynamic spectrum access Spectrum sensing Hidden Markov model Polynomial classifier Nonlinear autoregressive with exogenous inputs model Cognitive radio networks Mathematical models |
| url | http://hdl.handle.net/11073/7854 |