Receiver Artificial Noise Aided Terahertz Secure Communications with Eavesdropper in Close Proximity

Weijun Gao,Chong Han,Zhi Chen

DOI: https://doi.org/10.1109/globecom42002.2020.9322487

2020-01-01

Abstract:Although with narrow-beam transmissions, physical layer security of Terahertz (THz) communications faces great challenges with the presence of eavesdroppers in close proximity inside the Terahertz beam sector. This paper proposes a novel self-interference cancellation (SIC)-free receiver artificial noise (AN) assisted mechanism to address this challenge for THz secure communications. By exploiting the different temporal broadening effects of the artificial noise signals at the receiver side versus the eavesdropper side, the high-complexity SIC part can be mitigated with a proper receiver design. Moreover, the system parameters including power allocation, carrier frequency, transmit power, symbol time, pulse waveform type, and receiver parameter that maximize the secrecy rate of the THz communication system are solved, based on an efficient deep neural network (DNN) algorithm. Numerical results demonstrate that our proposed SIC-free receiver artificial noise scheme achieves 4 bps/Hz secrecy rate, with substantially lower hardware complexity than SIC-based receiver AN systems and reduced computational complexity than exhaustive search.