Resource Optimization in H-CRN with Supervised Learning Based Spectrum Prediction Technique

Abstract

Cognitive radio network shows potential means of granting intensifying demand for wireless applications. In this model, an efficient resource optimization scheme with Priority Pricing Technique (PPT) is proposed with supervised learning-based SVM to tackle limited spectrum availability and underutilization in Hybrid-Cognitive Radio Networks (H-CRN). H-CRN works under the principle of detection of PUs states (active/inactive). If spectrum sensing is made in favor of active PUs, then the CSI (Channel State Information) is estimated and works in underlay principle. If it is made in favor of inactive PUs, then the transmission is performed in overlay manner. In the proposed PPT the PUs and SUs with highest channel gain have the highest priority to use the spectral resources. SVM is used as an effective technique of spectrum sensing to provide higher probability of detection of PUs as soon as possible. The proposed method faces following challenges such as in order to enhance the CRN transmission performance, the PUs have to withstand more interference power and transmit power control is needed in improving the sum rates when the interference is severe in H-CRN. With Simulation outcomes, the assessment of the proposed (PPT) model among (Fixed Pricing Technique and Without Pricing Technique) indicates the proposed method's improved efficiency. The results reveal significant effectiveness in obtaining better classification accuracy with less computation complexity, increased throughput, spectral efficiency and energy efficiency of the network.