An Artificial-Noise-Aided Secure Scheme for Hybrid Parallel PLC/Wireless OFDM Systems

TL;DR

This paper proposes an artificial-noise-aided security scheme for hybrid PLC/wireless OFDM systems that enhances physical-layer security without needing eavesdropper channel information, by exploiting the media's decoupled nature.

Contribution

It introduces a novel AN-based security scheme for hybrid PLC/wireless OFDM systems that does not require eavesdropper channel state information and analyzes power allocation effects.

Findings

The scheme improves secure throughput against eavesdropping.

Power allocation significantly impacts security performance.

Effective in both single-link and two-link eavesdropping scenarios.

Abstract

We investigate the physical-layer security of indoor hybrid parallel power-line/wireless orthogonal-frequency division-multiplexing (OFDM) communication systems. We propose an artificial-noise (AN) aided scheme to enhance the system's security in the presence of an eavesdropper by exploiting the decoupled nature of the power-line and wireless communication media. The proposed scheme does not require the instantaneous channel state information of the eavesdropper's links to be known at the legitimate nodes. In our proposed scheme, the legitimate transmitter (Alice) and the legitimate receiver (Bob) cooperate to secure the hybrid system where an AN signal is shared from Bob to Alice on the link with the lower channel-to-noise ratio (CNR) while the information stream in addition to a noisy-amplified version of the received AN signal is transmitted from Alice to Bob on the link with higher CNR at each OFDM sub-channel. In addition, we investigate the effect of the transmit power levels at both Alice and Bob and the power allocation ratio between the data and AN signals at Alice on the secure throughput. We investigate both single-link eavesdropping attacks, where only one link is exposed to eavesdropping attacks, and two-link eavesdropping attacks, where the two links are exposed to eavesdropping attacks.