Outage reduction in 5G uplink cooperative NOMA using PSO-optimized relay selection and successive interference cancelation

Abstract

5G radio access networks face significant challenges, particularly related to user outages. While non-orthogonal multiple access (NOMA) has emerged as a promising solution for efficient spectrum sharing in 5G, it struggles to address outage issues. Reducing outage probability and enhancing system performance in NOMA networks is crucial. Previous efforts have focused on selecting cooperative relay nodes to mitigate outages, but most approaches fail to ensure low outage probability under adverse channel conditions. To address this, we propose a novel uplink cooperative NOMA system that utilizes successive interference cancelation (SIC) at the receiver end. In this system, a user with superior channel conditions serves as a relay for another user, ensuring more robust communication. Additionally, we integrate particle swarm optimization (PSO) for dynamic relay selection, effectively reducing outages, optimizing energy consumption, and significantly improving data rates. The inclusion of PSO-based relay selection demonstrates a substantial improvement in network performance. We compared the performance of particle swarm optimization (PSO) with that of the genetic algorithm (GA) and deep reinforcement learning (DRL). Our findings show that PSO achieved a throughput improvement rate of 68.76%, while GA resulted in an average improvement rate of 50.15%, and DRL produced a 46.38% improvement. These results highlight the effectiveness of PSO in enhancing system performance during outages. Furthermore, while PSO-based relay selection excels in outage scenarios, direct transmission proves more efficient under non-outage conditions, as revealed by our comparative analysis. This dual insight into relay and non-relay conditions provides a comprehensive understanding of the advantages and limitations of relay-based communication in 5G NOMA networks. However, instances of premature convergence were observed, warranting further investigation.