Cognitive radio (CR) is an enabling technology to effectively address the spectrum scarcity, and it will considerably
improve the spectrum utilization of future wireless communication systems. Cooperative spectrum sensing (CSS) has been
developed to take advantage of spectrum diversity of CR users. Security is an important issue, but not well addressed in CR
networks. This study considers a security problem arising from primary user emulation attack (PUEA) in CR networks. In
PUEA, a malicious attacker transmits an emulated primary signal and defrauds the CR users to prevent them from
accessing spectrum holes. We assume a smart and probabilistic attacker, in the sense that it is able to exactly detect the
spectrum holes and it transmits its signals in both vacant and occupied primary bands. Two important PUEA parameters are
estimated and then used to obtain the optimal voting rule for minimizing the global error probability. These attack
parameters include the probabilities of the PUEA signals in both occupied and unoccupied frequency bands. We formulate
and investigate the optimal voting rule for a CR network operating in the presence of a malicious attacker. The obtained
results indicate performance improvement provided by the proposed method compared with the conventional methods.