Reuse of Byzantine data in cooperative spectrum sensing using sequential detection

Abstract

Cooperative spectrum sensing (CSS) by exploiting diversity via the observations of spatially located secondary users improves the accuracy of the primary user (PU) detection, but cooperative paradigms are threatened by Byzantine attack. In this study, the authors propose a flexible Byzantine attack model, which goes beyond the existing models for its generalisation. Under this generalised Byzantine attack model, they give insights into the blind scenario where Byzantines make the fusion centre (FC) incapable of deciding the presence of the PU. To solve the blind problem, they formulate data transmission revelation (DTR) as trust reputation management to check consistency of the local decision. Moreover, they evaluate the usability of Byzantine data based on DTR and propose a sequential detection (SD) approach to reuse Byzantine data, which is a remarkable issue involved in CSS, however, ignored by most previous studies. Simulation results clearly reveal that in contrast to other approaches associated with sequential probability ratio test, the proposed SD benefits from Byzantine data to greatly improve the correct sensing ratio and the sample size, and still functions well in the blind scenario.