Average Block Error Rate of Intelligent Reflecting Surface Aided Short-Packet Communications

Abstract

<p>We comprehensively analyze the average block error rate (BLER) performance of intelligent reflecting surface (IRS) aided short-packet communication (SPC) systems over Rician fading. Our analysis covers two communication scenarios, depending on whether or not the direct link exists between the base station and the user, addresses two types of phase errors, namely, phase estimation errors modeled by the Von Mises distribution and quantization errors, and facilitates system design based on the instantaneous channel state information (CSI) and the statistical CSI. In our analysis, we first derive the distributions of the end-to-end channel and the received signal-to-noise ratio at the user under different phase adjustment strategies. Then we propose a more accurate approximation of the $Q$ function to derive closed-form expressions for the average BLER. Using numerical and simulation results, we corroborate the accuracy of our analysis and investigate the impacts of various parameters of the system, such as the number of IRS elements and the phase adjustment strategy, and different CSI availabilities on the average BLER performance. We also show that compared to the instantaneous CSI based design, the statistical CSI based design yields the near-optimal BLER performance for a large Rician factor, making it attractive for ultra-reliable and low-latency communications, given its low implementation complexity.</p>