An overview of Cooperative Spectrum Sensing based on Machine Learning Techniques

Cognitive Radio is an intelligent wireless communication system able of learning from the environment. It allows reusing of the radio resources available by users called Secondary Users without creating harmful interference with licensed users. Spectrum Sensing is the important function in Cognitive Radio technology. The application of Machine Learning techniques for Spectrum Sensing has a drawn interest in the literature. In this paper, we present an overview on Machine Learning techniques for Cooperative Spectrum Sensing. We compare three Machine Learning techniques, (Naive Bayes, SVM and Multilayer Perceptron) with traditional cooperative Spectrum Sensing techniques (Maximum Ratio Combining MRC, OR rule and AND rule), under AWGN channel and Rayleigh channel. The performance evaluated using ROC curve, and the results demonstrate that MRC is the optimum technique under AWGN channel and SVM with Gaussian kernel function is the optimum Machine Learning technique under Rayleigh channel.