IMPROVING THE QUALITY OF SERVICE OF A COGNITIVE RADIO NETWORK THROUGH SPECTRUM SENSOR OPTIMIZATION

Subject Area: Computer Science

The underutilization of the constrained spectrum by the fixed channel allocation strategy and the increasing demand for seamless wireless services provide several challenges for wireless communications. Mobile users need to create appropriate channel allocation algorithms for continuous communication in order to improve spectral efficiency. The Cognitive Radio Network (CRN) offers a solution to the inherent spectrum scarcity problem in 4G and other networks by allowing unlicensed Secondary Users (SUs) with cognitive devices to utilise available spectrum when licenced Primary Users (PUs) are not using it. This is achieved through dynamic spectrum access. SU measurements made during the sensing process are sometimes unclear because to channel conditions that fluctuate, multipath fading, and shadowing.This causes the SUs to make incorrect choices about switching, resulting in continuous spectrum handoff and the undesirable ping-pong effect. The Support Vector Machine (SVM) classifier proposed in this study splits the spectrum into two categories: busy and idle. Then, using an underlay spectrum access model, a QoS-aware Adaptive Neuro-Fuzzy Inference System (ANFIS) framework is created for spectrum switching decisions. It is based on the fluctuating channel state information, the dynamic activities of PUs, and the heterogeneous Quality of Service (QoS) requirements of the SUs. The present study included three distinct techniques, including the qualitative, experimental, and simulation approaches.The primary finding indicates that adopting a proactive approach during the spectrum decision and channel classification phases enhances the effectiveness of a cognitive network by decreasing the amount of time required to allocate spectrum resources to service units (SUs). This, in turn, lowers the frequency of collisions, resulting in more effective data transfer.