Toma, Ogeen 
ORCID: 0000-0002-3969-0470
(2022)
Channel Activity Statistics Estimation in Spectrum Sharing Systems Based on Imperfect Spectrum Sensing.
PhD thesis, University of Liverpool.
|
Text
201300985_May2022.pdf - Unspecified Download (10MB) | Preview |
Abstract
As we are stepping into the era of beyond 5G, the demand for frequency bands will increase significantly to accommodate the fast growing tendency in wireless communications technology. Spectrum sharing is one of the promising solutions to overcome the frequency scarcity problem and maximise spectrum utilisation efficiency. Its consideration can be seen in the recent ongoing deployment of 5G as in 5G New Radio Unlicensed (5G NR-U). The harmonious coexistence of several wireless systems in a shared frequency spectrum is highly dependent on making effective decisions for the utilisation of such spectrum. These decisions are usually based on the users’ traffic activity within the channel and their statistical information. Therefore, it is crucial for a spectrum sharing system to accurately obtain channel activity statistics. Although spectrum sensing is used in such systems to determine the instantaneous state of the channel, sensing decisions can further be exploited to provide a broad range of statistical information of the channel activity. However, spectrum sensing is imperfect in real world which therefore leads to inaccurate estimation of the channel activity statistics. In this context, this thesis studies the problem of estimating the channel activity statistics under (realistic) Imperfect Spectrum Sensing (ISS) and it finds mathematical relationships (in closed-form expressions) between the observed channel activity statistics under ISS and their corresponding actual statistical information, as a function of relevant operating parameters including the probability of sensing errors, the employed sensing period and the sample size. Such problem is poorly addressed in the literature, without deep and rigorous mathematical analyses taking into account all the factors that would influence the estimation accuracy of the channel activity statistics. Then, the thesis investigates different approaches that can be used to improve the estimation of the channel activity statistics under ISS, namely the closed-form expression approach, which is based on the obtained mathematical expressions for these statistics; the algorithmic approach, which is based on reconstruction algorithms; and finally a Traffic Learning (TL) approach, which is based on deep learning techniques. It is shown that the proposed estimation methods in this thesis outperform the existing methods in the literature without requiring any prior knowledge of the channel activity. The correctness of the obtained analytical expressions and proposed estimation methods are corroborated with both simulation and experimental results, for which a USRP-based prototype is developed as an experimental platform to validate the theoretical analyses conducted for the estimation of the channel activity statistics in spectrum sharing systems.
| Item Type: | Thesis (PhD) |
|---|---|
| Divisions: | Faculty of Science and Engineering > School of Electrical Engineering, Electronics and Computer Science |
| Depositing User: | Symplectic Admin |
| Date Deposited: | 08 Jun 2022 08:55 |
| Last Modified: | 18 Jan 2023 21:00 |
| DOI: | 10.17638/03155667 |
| Supervisors: |
|
| URI: | https://livrepository.liverpool.ac.uk/id/eprint/3155667 |
Dimensions
Dimensions
