Online Malicious Behavior Detection in Collaborative Spectrum Sensing: A Change Detection Approach

Intelligent attackers in a collaborative spectrum sensing system could act as honest users to conceal themselves and start malicious behavior abruptly since an unpredictable time slot. Affected by honest behavior before attacking time, traditional malicious behavior detection (MBD) algorithms are not agile enough to identify the abrupt change of behavior. To alleviate this challenge, in this paper, we propose the Rao test-based malicious behavior detection (RTMBD) algorithm, which could detect the malicious behavior with unknown parameter and unknown starting time. The proposed RT-MBD is not affected by honest behavior before attacking time and has a shorter detection delay with constraint of a certain false alarm rate than conventional algorithms. Performance of RT-MBD is validated by both mathematical proof and numerical experiments.