Spectrum sensing under distribution uncertainty in cognitive radio networks

Abstract

The successful coexistence of cognitive radio systems with licensed system requires the secondary users the capability of interference-awareness, i.e., knowing which spectrum bands are occupied by primary users, i.e., the legacy users. Spectrum sensing thus is a key enabling module, which usually models the sensing process as a binary hypothesis testing assuming known signal distribution. However, an unrealistic assumption regarding the signal distribution easily leads to unreliable detection probability. In this paper, we study the sensing performance considering the distribution uncertainty in hypothesis testing, i.e., the actual distribution function of the received signal strength is not known. According to different signal characteristics, we define appropriate uncertainty sets respectively for different hypotheses. Then we present an approximate approach to determine the robust decision threshold, and investigate the performance bounds for the detection probability under distribution uncertainty. Moreover, we provide an analytical expression for the lower bound of detection probability. Numerical results are given to validate our conclusions.