Faculty Publications

Title

A frequency domain two-stage spectrum sensing method based on SVM

Document Type

Conference

Keywords

frequency domain detection, noise estimation, spectrum sensing, support vector machine, two-stage sensing

Journal/Book/Conference Title

IEEE International Conference on Electro Information Technology

First Page

105

Last Page

108

Abstract

Spectrum sensing is one of the key technologies of cognitive radio. Based on noise characteristics estimation and support vector machine (SVM) technology, this paper proposed a frequency domain two-stage spectrum sensing method to improve sensing accuracy under low signal-to-noise ratio scenarios with low system complexity and high generalization ability. In the slow sensing stage, the frequency-domain data is used as the feature data of the SVM and the noise parameters (the variance and the mean) are estimated. In the fast sensing stage, only the variance and the mean of the input signals are used as the feature data of the SVM. The simulation results show that the average detection error probability of the proposed method could be as low as 5% when the signal-to-noise ratio is -15dB.

Original Publication Date

9-27-2017

DOI of published version

10.1109/EIT.2017.8053339

Repository

UNI ScholarWorks, Rod Library, University of Northern Iowa

Language

en

Share

COinS