A Cooperative Spectrum Sensing Architecture and Algorithm for Cloud- and Big Data-based Cognitive Radio Networks

Cognitive Radio Network (CRN) was designed to lessen the shortage of radio resources. The Secondary Users (SUs) can opportunistically utilize any available spectrum when the Primary Users (PUs) are inactive. Some of the challenges of CRN include the service interruption loss, complexity of processing and exchange of large amount of data, limited available memory to SUs and the non-real-time exchange of spectrum sensing data. These challenges can lead to significant degradation in the performance of a CRN. Therefore, there is a need to seek solutions that will alleviate these problems. The Cloud system incorporated with Big Data Analytics algorithm can be a potential solution. In this paper, we propose a Cloud-based Cooperative Spectrum Sensing model for CRN that allows the SUs to aggregate their individual spectrum sensing data into a cloud environment, where it can be analyzed using a proposed expanded Apache Spark algorithm incorporated with the hybridization of three machine learning methods-ensemble classifier approach that can effectively and efficiently analyze the spectrum sensing data for easy access, real-time analysis, deep insight and on-demand decision support for the SUs. In addition, the two-layer Fusion Center design proposed introduces redundancy by using the cloud as a secondary Fusion Center while still maintaining a primary land-based Fusion Center.