Artificially Noise-Injected Low-Density Parity-Check Codes for the Gaussian Wiretap Channel

Abstract

Wireless telemetry communication systems may be vulnerable to eavesdroppers due to their broadcasting nature, which is a risk to confidential information transmission. This paper explores the feasibility of low-density parity-check (LDPC) codes for wiretap coding over a Gaussian channel for standard telemetry links. By introducing a wiretap LDPC coding method, we aim to mitigate the risk of eavesdropping. We further explore the notion of fine-tuning the trade-off in these codes between secrecy and reliability through artificial noise injection. Simulation results demonstrate the effectiveness of employing an LDPC-based wiretap code over telemetry links to provide secrecy. The analysis of the code shows that a neural network-based mutual information estimator can be utilized to calculate information leakage over telemetry links.