Statistical-based detection of pilot contamination attack for NOMA in 5G networks

Fifth-generation (5G) communication technologies, such as millimeter wave communication, massive multiple-input-multiple-output and non-orthogonal-multiple-access (NOMA) are playing a pivotal role in promoting the modern applications of the Internet-of-Things. Using non-orthogonal resource allocation, NOMA can increase spectrum efficiency and achieve wide connectivity with low transmission delay and signaling cost. Despite the high potential of NOMA in 5G communications, NOMA is susceptible to a pilot contamination attack (PCA), in which an attacker resents the same pilot signals as authorized users. Currently, using the available detection methods in NOMA gives high false positive probability since the time-division-duplex or orthogonal resource block can be allocated by many authorized user. Since the pilot contamination attack changes the signal reception at the legitimate receiver, this work introduces a novel detection scheme for identifying Pilot Contamination attack (PCA) that statistically investigates the asymmetry in received signal power levels. The main idea of the proposed detection scheme is to use various statistical measurements for normal traffic attributes (CSI) as a reference profile. Then, compute the Mahalanobis distance between the reference profile and CSI for the incoming connection and use the probability of the uniform distribution to make the final detection decision. The performance of the proposed detection technique in terms of its detection rate and false positive probabilities has been evaluated through extensive simulation. The simulation results show that the proposed scheme succeeded in detecting the pilot contamination attack with a detection rate of up to 98% and a precision reached 97.88%.