Channel Decontamination Based Pilot Contamination Attack Resistance for mmWave Grant-Free Massive MTC Networks: A Distributed Compressive Sensing Approach with Multidimensional Dictionaries

Due to the lack of authentication, millimeter-wave (mmWave) grant-free (GF) massive machine-type communication (mMTC) access networks are vulnerable to pilot contamination attack (PCA), which will cause serious degradation of channel estimation and disrupt data reception. However, existing PCA defenses for mmWave GF-mMTC networks mainly detect attackers without recovering data from legitimate users (LUs). To solve this problem, we propose a channel decontamination based PCA resistance scheme for mmWave GF-mMTC networks to jointly detect attackers, decontaminate attacked LU channels, and recover legitimate data. We first establish an angular-delay-slot domain channel model that enhances the sparsity of each transmitter’s channel virtual representation (CVR), in order to reduce CVR overlap and facilitate channel decontamination. Based on the established channel model, we formulate the PCA resistance problem as a distributed compressive sensing problem aiming to reconstruct the CVR of each transmitter with angulardelay joint sparsity in successive time slots. Then, we develop a multidimensional dictionary based inner-outer support matching pursuit (MD-IOSMP) algorithm that exploits temporal LU support correlation for reliable CVR reconstruction and channel decontamination under the quasi-static scenario. Moreover, to accommodate practical dynamics, which may partially violate the temporal correlation assumption, we further develop a dynamicaware (DA) version of the MD-IOSMP algorithm, termed MD-IOSMP- DA, which applies a tolerance-based criterion to identify legitimate channel components to maintain system performance reliability under dynamic propagation conditions. Simulation results show that the proposed scheme can achieve a substantial performance gain over several reference schemes.