Extreme Learning Machine Based Cooperative Spectrum Sensing in Cognitive Radio Networks

Abstract

In this paper, we consider the application of Extreme Learning Machine (ELM) with Neural Networks in cooperative spectrum sensing (CSS) for cognitive radio networks (CRN). Based on a statistical analysis of classical energy detector, the probability of detection and the false alarm has been calculated, which depends solely on SNR and the number of samples values. The channel occupancy detection results obtained from the proposed approach are compared with established analytical techniques such as MRC and AND/OR rules and well-known Machine Learning (ML) techniques, including, Support Vector Machine (SVM) and K-Means. The Comparison matrices were receiver operating characteristic (ROC) curve and area under the curve (AUC). We obtain the computational performance of the aforementioned NNELM model during the training phase and calculated the channel detection probability. Ultimately, the results demonstrate that the NN-ELM technique presents a better trade-off between training time and detection performance.