Short-Packet URLLCs for Multihop MIMO Full-Duplex Relay Networks: Analytical and Deep-Learning-Based Real-Time Evaluation

Abstract

<p>In this study, we investigate multihop multiple-input multiple-output (MIMO) full-duplex relay (FDR) networks with short-packet ultra-reliability and low-latency communications (uRLLCs), where transmit-antenna selection (TAS)/maximum-ratio combining (MRC), TAS/selection combining, and maximum-ratio transmission/MRC are leveraged as diversity techniques.</p> <p>Under quasi-static Rayleigh-fading channels and the finite-blocklength theory, the end-to-end block-error rate (BLER) performance of the considered FDR is analyzed and compared with that of half-duplex relaying systems, from which the effective throughput (ETP), energy efficiency (EE), reliability, and latency are also investigated. To gain further insights into the system design, an asymptotic BLER analysis in the high signal-to-noise ratio regime is provided. Based on the analytical and simulation results, the gains of multihop MIMO FDR networks in the context of short-packet uRLLCs are confirmed via the effect of system parameters on network performance. In addition, a novel deep multiple-output neural-network framework is developed to simultaneously predict the ETP, EE, and reliability with high accuracy, low computational complexity, and short execution time. This makes the framework an efficient estimator for real-time evaluations of future IoT applications, compared with conventional analysis and simulation approaches.</p>