Performance Evaluation of Simultaneously Transmitting and Reflecting Intelligent Reflecting Surface - Assisted NOMA Networks for Short-Packet Communications

Abstract

In this paper, we investigate a non-orthogonal multiple access network supported by both reflecting surfaces and simultaneous transmission (STAR-IRS-NOMA) for short-packet communication. Firstly, the performance metrics of the STAR-IRS-NOMA network, including block error rate (BLER), delay, and throughput for each user, are devired over Rayleigh fading channels. Subsequently, based on Monte Carlo simulations, we verify the accuracy of the derived expressions and demonstrate the superior performance of the STAR-IRS-NOMA system compared to the conventional system (IRS-NOMA). Finally, a deep neural network (DNN) is proposed to predict the delay and throughput of the STAR-IRS-NOMA system. The predictive results provide evidence of the DNN's exceptional advancement in achieving short processing times and high accuracy compared to alternative methods.