Spectrum Sensing and Attack Node Detection for Cognitive Radios based on RVM

Abstract

In a complex geographical environment, communication quality of communication equipment is being seriously challenged. Secondary Users(SUs) must make the best possible use the idle spectrums that Primary Users(PUs) do not use and change spectrum frequently. Using the relevance vector machine(RVM) to establish a signal noise Ratio(SNR) Model for interference information and bit error rate(BER). Through the model and real-time interference information, the minimum channel SNR meeting the BER requirements of communication equipment can be predicted. According to the simulation results, this method has better performance for selecting available channel and detecting attack node.