HMM-Based Malicious User Detection for Robust Collaborative Spectrum Sensing

摘要

Collaborative spectrum sensing improves the spectrum state estimation accuracy but is vulnerable to the potential attacks from malicious secondary cognitive radio (CR) users, and thus raises security concerns. One promising malicious user detection method is to identify their abnormal statistical spectrum sensing behaviors. From this angle, two hidden Markov models (HMMs) corresponding to honest and malicious users respectively are adopted in this paper to characterize their different sensing behaviors, and malicious user detection is achieved via detecting the difference in the corresponding HMM parameters. To obtain the HMM estimates, an effective inference algorithm that can simultaneously estimate two HMMs without requiring separated training sequences is also developed. By using these estimates, high malicious user detection accuracy can be achieved at the fusion center, leading to more robust and reliable collaborative spectrum sensing performance (substantially enlarged operational regions) in the presence of malicious users, as compared to the baseline approaches. Different fusion methods are also discussed and compared.