Toward Secure Centralized Spectrum Sensing by Utilizing Geographical Information

Abstract

Spectrum utilization becomes more and more important while new techniques keep increasing and the spectrum bands remain finite. Cognitive Radio (CR) is a revolutionary technology to make use of the spectrum more effectively. In order to avoid the interference to the primary user, spectrum sensing must be sensitive. Cooperative Spectrum Sensing (CSS) is one way to increase the reliability of spectrum sensing. The information fusion technique is a key component of CSS. In this paper, we adopt a grid-like model for CR networks, and we utilize geographical information with reputational weights to propose a two-level fusion scheme called Secure Centralized Spectrum Sensing (SCSS). The simulation results show that as the attackers present high density aggregation at some areas, the correct sensing ratio of SCSS is increasing as well even when the number of attackers is very large.