The inherent nature of cognitive radio (CR)
networks has brought new threats to wireless communications. Primary user emulation attack (PUEA) has been
widely studied as a serious threat to cooperative spectrum
sensing (CSS) in CR networks. In PUEA, a malicious user
can obstruct CR users from accessing idle frequency bands
by imitating licensed primary user (PU) signal characteristics. The present study introduces a new CSS scheme in
the presence of a malicious PUEA based on multi-level
hypothesis testing (MLHT). In the proposed method, generalizing from binary hypothesis testing to MLHT, we
partition the decision space to four decision options and
apply minimum Bayes cost criteria to determine the
channel status. We also discuss practical limitation issues
that need to be considered when applying the MLHT
approach. Simulation results are provided to indicate the
performance improvement of the proposed MLHT method
against PUEA, compared with the conventional method.