Secrecy Rate Analysis and Active Pilot Attack Detection for IRS-Aided Massive MIMO Systems

The active pilot contamination attacks in intelligent reflecting surface (IRS) aided massive multiple-input multiple-output systems are investigated. By proposing a statistical channel state information based IRS phase-shift optimization technique, an achievable secrecy rate is derived in the presence of practical impediments, including erroneously estimated composite channels via linear minimum mean square error estimation criterion, residual interference due to active pilot contamination, artificial noise (AN) generation, and spatially correlated fading at the base-station antennas and IRS elements. A transmit power allocation technique is also proposed. Two active pilot attack detectors are designed based on the Neyman-Pearson and generalized likelihood ratio test criteria. The performance of these detectors is investigated by deriving the probability of detection, probability of false alarm, and receiver operating characteristics. Our secrecy rate analysis reveals that the rate leaked into the eavesdroppers by active pilot contamination attacks can be considerably high. The proposed power allocation algorithm jointly assigns transmit powers for the legitimate signals and AN sequences for maximizing the minimum secrecy rate of the weakest legitimate user to ensure user-fairness. The proposed detectors of active pilot attacks may be useful in designing remedial techniques to mitigate detrimental effects of active eavesdropping.