Analysis and implementation of a wavelet based spectrum sensing method for low SNR scenarios

In Cognitive Radio applications, spectrum sensing plays a fundamental role in order to learn the behavior of primary users (PUs) and access to the spectral resource opportunistically. Among the available methods, a surely promising approach is the wavelet based one. It allows to subdivide the wide-band spectrum under analysis in a proper number of sub-bands, based on Power Spectral Density (PSD) irregularities, remarked by the extrema of the Continuous Wavelet Transform (CWT) first derivative. Generally, such kind of methods works well as long as good Signal-to-Noise Ratio (SNR) can be experienced over the span of interest. In this context, starting from an approach present in literature, the present work proposes, customizes and implements a wavelet based spectrum sensing method, thought to operate also in challenging SNR scenarios.