High-energy-efficiency spectrum sensing method based on support vector machine (SVM)

Abstract

The invention provides a high-energy-efficiency spectrum sensing method based on a support vector machine (SVM). Firstly, energy statistics amounts of a main user signal and a noise are calculated, energy vectors are obtained, and then residual vectors of the main user signal and the noise are obtained; secondly, an SVM classification model is trained offline by utilizing an eigenvector, and Gaussian kernel parameters are optimized on the basis of simulated annealing; and finally, characteristic expressions of received signals are calculated, classification is carried out by using the SVM classification model, and whether the signals include a main user or not is judged so as to judge a frequency spectrum state.

Description

technical field [0001] The invention relates to spectrum sensing technology in the field of cognitive radio, in particular to an energy-efficient spectrum sensing method based on a support vector machine. Background technique [0002] Usually, a spectrum is only allocated to primary users (PUs) and perceived users cannot access this frequency band. However, the primary user does not occupy this frequency band all the time, so the idle frequency band leads to lower spectrum utilization. Spectrum sensing can detect unoccupied frequency bands, so that sensing users can use this frequency band when the primary user is not occupied, improving spectrum utilization. [0003] The spectrum sensing algorithm based on energy detection is the most commonly used sensing model. This scheme does not require obtaining prior knowledge about the primary user signal. Researchers have successively proposed single-threshold and double-threshold spectrum sensing algorithms. In order to furthe...

AI Technical Summary

Problems solved by technology

However, when the signal-to-noise ratio is very low and the number of sampling point

Images