Partial and Full Relay Selection Algorithms for AF Multi-Relay Full-Duplex Networks With Self-Energy Recycling in Non-Identically Distributed Fading Channels

• Published in: IEEE transactions on vehicular technology Vol. 71; no. 6; pp. 6173 - 6188
• Main Authors: Nguyen, Tan N., Duy, Tran Trung, Tran, Phuong T., Voznak, Miroslav, Li, Xingwang, Poor, H. Vincent
• Format: Journal Article
• Language: English
• Published: New York IEEE 01.06.2022; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Algorithms; Channels; full relay selection; Full-duplex; Interference; Interference cancellation; Monte Carlo simulation; Networks; NOMA; non-identically distributed; outage probability; Outages; partial relay selection; Protocols; Random variables; Recycling; Relay networks; Relay networks (telecommunication); Relays; self-energy recycling; Signal to noise ratio; Wireless communication
• ISSN: 0018-9545; 1939-9359
• DOI: 10.1109/TVT.2022.3158340

Full-duplex communication offers enhanced spectral efficiency for relay deployment, but suffers from the inherent self-interference from the strong transmit signal coupling to the sensitive receive chain. In this article, we propose a self-energy recycling (S-ER) protocol for full-duplex multi-relay networks, in which the energy from self-interference is harvested back at the relay for future use. Furthermore, two amplify-and-forward (AF) relay selection algorithms, namely, partial relay selection (PRS) and full relay selection (FRS) are introduced to enhance the reliability of the proposed systems. For PRS, the best relay is selected based on just the knowledge of the channels from the source to all relays, while in FRS, the best relay is selected based on the end-to-end signal-to-noise ratio, which requires knowledge of all source-relay and relay-destination links. We provide a thorough analysis on the outage performance and the spectrum efficiency of the proposed algorithms in both cases: all channel gains are independently but non-identically distributed (i.n.d.) (case 1) or independently, identically distributed (i.i.d.) (case 2) Rayleigh random variables. It is shown that SER and FRS can significantly enhance the performance of FD networks and avoid the outage floor when the number of relays increases, while the outage probability (OP) in PRS case reaches an outage floor. In addition, the end-to-end signal-to-noise ratio in both cases can be minimized if an optimal power-splitting factor is selected. All analytical results are verified by Monte Carlo simulation.