Improper Gaussian signaling for multiple-access channels in underlay cognitive radio

TL;DR

This paper explores the use of improper Gaussian signaling in a multiple-access channel within underlay cognitive radio, demonstrating its optimality under certain interference conditions and providing an algorithm for optimal transmission parameter computation.

Contribution

It introduces the concept of improper Gaussian signaling for secondary users in underlay cognitive radio, showing its optimality and developing an efficient algorithm for rate region boundary attainment.

Findings

Improper signaling is optimal when interference exceeds a threshold.

An efficient algorithm for transmission parameter optimization is proposed.

Improper Gaussian signaling enhances secondary user performance in cognitive radio.

Abstract

This paper considers an unlicensed multiple-access channel (MAC) that coexists with a licensed point-to-point user, following the underlay cognitive radio paradigm. We assume that every transceiver except the secondary base station has one antenna, and that the primary user (PU) is protected by a minimum rate constraint. In contrast to the conventional assumption of proper Gaussian signaling, we allow the secondary users (SUs) to transmit improper Gaussian signals, which are correlated with their complex conjugate. When the secondary base station performs zero-forcing, we show that improper signaling is optimal if the sum of the interference channel gains (in an equivalent canonical model) is above a certain threshold. Additionally, we derive an efficient algorithm to compute the transmission parameters that attain the rate region boundary for this scenario. The proposed algorithm exploits a single-user representation of the secondary MAC along with new results on the optimality of improper signaling in the single-user case when the PU is corrupted by improper noise.