Publicação
Wideband spectrum sensing for dynamic spectrum sharing
| Resumo: | The proliferation of wireless devices grows exponentially, demanding more and more data communication capacity over wireless links. Radio spectrum is a scarce resource, and traditional wireless networks deployed by Mobile Network Operators (MNO) are based on an exclusive spectrum band allocation. However, underutilization of some licensed bands in time and geographic domains has been reported, especially in rural areas or areas away from high population density zones. This coexistence of increasingly high data communication needs and spectrum underutilization is an incomprehensible scenario. A more rational and efficient use of the spectrum is the possibility of Licensed Users (known as Primary Users – PU) to lease the spectrum, when not in use, to Unlicensed Users (known as Secondary Users – SU), or allowing the SU to opportunistically use the spectrum after sensing and verifying that the PU is idle. In this latter case, the SU must stop transmitting when the PU becomes active. This thesis addresses the spectrum sensing task, which is essential to provide dynamic spectrum sharing between PUs and SUs. We show that the Spectral Correlation Function (SCF) and the Spectral Coherence Function (SCoF) can provide a robust signal detection algorithm by exploiting the cyclostationary characteristics of the data communication signal. We enhance the most used algorithm to compute de SCF - the FAM (FFT Accumulation Method) algorithm – to efficiently compute the SCF in a local/zoomed region of the support ( ; ) plane (frequency/cycle frequency plane). This will provide the quick identification of spectral bands in use by PUs or free, in a wideband sampling scenario. Further, the characterization of the probability density of the estimates of the SCF and SCoF when only noise is present, using the FAM algorithm, will allow the definition of an adaptive threshold to develop a blind (with respect to the noise statistics) Constant False Alarm Rate (CFAR) detector (using the SCoF) and also a CFAR and a Constant Detection Rate (CDR) detector when that characterization is used to obtain an estimate of the background noise variance (using the SCF). |
|---|---|
| Autores principais: | Tavares, Luís Miguel |
| Assunto: | Spectrum sensing Spectrum sharing 5G Cognitive radio Cyclostationarity FAM - FFT Accumulation Method |
| Ano: | 2022 |
| País: | Portugal |
| Tipo de documento: | tese de doutoramento |
| Tipo de acesso: | acesso aberto |
| Instituição associada: | Instituto Politécnico de Beja |
| Idioma: | inglês |
| Origem: | Repositório Institucional do IPBeja |
| Resumo: | The proliferation of wireless devices grows exponentially, demanding more and more data communication capacity over wireless links. Radio spectrum is a scarce resource, and traditional wireless networks deployed by Mobile Network Operators (MNO) are based on an exclusive spectrum band allocation. However, underutilization of some licensed bands in time and geographic domains has been reported, especially in rural areas or areas away from high population density zones. This coexistence of increasingly high data communication needs and spectrum underutilization is an incomprehensible scenario. A more rational and efficient use of the spectrum is the possibility of Licensed Users (known as Primary Users – PU) to lease the spectrum, when not in use, to Unlicensed Users (known as Secondary Users – SU), or allowing the SU to opportunistically use the spectrum after sensing and verifying that the PU is idle. In this latter case, the SU must stop transmitting when the PU becomes active. This thesis addresses the spectrum sensing task, which is essential to provide dynamic spectrum sharing between PUs and SUs. We show that the Spectral Correlation Function (SCF) and the Spectral Coherence Function (SCoF) can provide a robust signal detection algorithm by exploiting the cyclostationary characteristics of the data communication signal. We enhance the most used algorithm to compute de SCF - the FAM (FFT Accumulation Method) algorithm – to efficiently compute the SCF in a local/zoomed region of the support ( ; ) plane (frequency/cycle frequency plane). This will provide the quick identification of spectral bands in use by PUs or free, in a wideband sampling scenario. Further, the characterization of the probability density of the estimates of the SCF and SCoF when only noise is present, using the FAM algorithm, will allow the definition of an adaptive threshold to develop a blind (with respect to the noise statistics) Constant False Alarm Rate (CFAR) detector (using the SCoF) and also a CFAR and a Constant Detection Rate (CDR) detector when that characterization is used to obtain an estimate of the background noise variance (using the SCF). |
|---|