Time-domain artificial noise generation technique using time-domain and frequency-domain processing for physical layer security in MIMO-OFDM systems

Abstract

Artificial noise (AN) is one of physical layer security techniques, which can realize secure communication. AN is a noise signal which is intentionally transmitted from transmitter along with data signals. Since AN is generated to be canceled out at legitimate receivers, it can prevent eavesdroppers from eavesdropping even if transmitters do not know the existence of eavesdroppers. In spite of its usefulness, most of the AN generation techniques cannot be applied to multiple-input multiple-output (MIMO) systems when the number of antennas at the transmitter is less than that at the legitimate receiver. In this paper, we consider time-domain AN generation techniques for MIMO orthogonal frequency division multiplexing (OFDM) systems. A time-domain AN generation technique is originally proposed for single-input single-output (SISO) OFDM systems. Thus, we first consider to apply the conventional time-domain AN generation technique to MIMO-OFDM systems, and show that this technique requires a large number of antennas at the transmitter and/or a longer cyclic prefix (CP). Next, to solve this problem, we propose another time-domain AN generation technique. Our proposed time-domain AN is generated to be canceled out at the legitimate receiver in the frequency domain, while the conventional time-domain AN is canceled out in the time domain. As a result, the proposed time-domain AN generation technique can be applied to MIMO-OFDM systems, irrespective of the number of antennas and the length of CP. Simulation results can show that the proposed AN achieves the same secrecy rate as the conventional time-domain AN. Moreover, the proposed technique achieves a higher secrecy rate when a MIMO-OFDM system is employed with the antenna configuration to which the conventional time-domain AN generation technique is hard to be applied.