Covert Transmission with a Self-sustained Relay

TL;DR

This paper investigates the limits and costs of covert communication via a self-sustained relay that harvests energy from the source, analyzing detection error probabilities and effective covert rates under different energy harvesting schemes.

Contribution

It derives the detection error probability and maximum covert rate for a self-sustained relay using time switching and power splitting schemes, highlighting their equivalence in detection performance.

Findings

Detection error probability is the same for TS and PS schemes.

Maximum covert rate depends on energy conversion efficiency and covertness constraints.

TS scheme can outperform PS in low transmit power scenarios.

Abstract

This work examines the possibility, performance limits, and associated costs for a self-sustained relay to transmit its own covert information to a destination on top of forwarding the source's information. Since the source provides energy to the relay for forwarding its information, the source does not allow the relay's covert transmission and is to detect it. Considering the time switching (TS) and power splitting (PS) schemes for energy harvesting, where all the harvested energy is used for transmission at the self-sustained relay, we derive the minimum detection error probability $\xi^{\ast}$ at the source, based on which we determine the maximum effective covert rate $\Psi^{\ast}$ subject to a given covertness constraint on $\xi^{\ast}$. Our analysis shows that $\xi^{\ast}$ is the same for the TS and PS schemes, which leads to the fact that the cost of achieving $\Psi^{\ast}$ in both the two schemes in terms of the required increase in the energy conversion efficiency at the relay is the same, although the values of $\Psi^{\ast}$ in these two schemes can be different in specific scenarios. For example, the TS scheme outperforms the PS scheme in terms of achieving a higher $\Psi^{\ast}$ when the transmit power at the source is relatively low. If the covertness constraint is tighter than a specific value, it is the covertness constraint that limits $\Psi^{\ast}$, and otherwise it is upper bound on the energy conversion efficiency that limits $\Psi^{\ast}$.