Optimal data fusion of collaborative spectrum sensing under attack in cognitive radio networks

Abstract

Cognitive radio networks allow opportunistic spectrum access and can significantly improve spectral efficiency. To achieve higher sensing accuracy, cognitive radio systems often require cooperation among secondary users. One of the most important aspects in collaborative spectrum sensing is the data fusion algorithm which combines the sensing results from secondary users to produce the final channel status hypothesis. However, plenty of factors may affect the performance of certain data fusion rule, for example, the individual sensing node's sensing accuracy, the number of involved nodes, and the like. If Spectrum Sensing Data Falsification (SSDF) attack exists, it will become more challenging to make proper data fusion. In this article, we first introduce framework, and then evaluate the data fusion rules in different scenarios through simulation examples. Finally, a Genetic Algorithm based optimal scheme is proposed to achieve better performance in all scenarios.