Performance of Spectrum Sensing in Different Fading Environments for Cognitive Radios

Radio spectrum is a very important natural resource and efficient utilization of spectrum increases the QOS of wireless communication systems. Cognitive Radio Technology plays a very important role for proper usage of spectrum having efficient algorithms for spectrum sensing. This study focuses more on Cooperative spectrum sensing in which decision made by fusing the spectrum sensing result of multiple secondary users in a system. Among the different methods of cooperative spectrum sensing like AND/OR, MRC method shows greater efficiency in spectrum sensing. But, Spectrum sensing is greatly influenced by the wireless channel parameters like path loss, shadowing and fading effects.Spectrum sensing can be done by using Machine Learning algorithms and it shows better spectrum sensing results in different fading environments. This study has performed spectrum sensing in different fading environments like Rayleigh, Rician and Nakagami channels, where the fusion result from all the cooperative secondary nodes is calculated by using Support Vector Machine Algorithm. If the different channel effects are considered, the SVM algorithm gave better results for spectrum sensing. Here, simulations are performed with different training data sizes and the performance is studied for each trial using the linear and Gaussian kernels of SVM algorithm.