Resource Allocation and Beamforming Design in the Short Blocklength Regime for URLLC

Providing ultra reliable and low-latency communication (URLLC) is considered one of the major challenges for wireless communication networks. This article considers a downlink URLLC system in which a base station (BS) serves multiple single-antenna users in the short blocklength regime. With the objective of maximizing the users' minimum rate, three different optimization problems are considered: (i) joint design of bandwidth and power allocation for the case of a single-antenna BS; (ii) beamforming design for the case of a multiple-antenna BS; and (iii) design of power allocation with regularized zero-forcing beamforming for the case of a multiple-antenna BS. In the short blocklength regime, the achievable rate is a complicated function of bandwidth and power allocation coefficients or beamforming vectors, which makes these max-min rate optimization problems challenging to solve. This work develops path-following algorithms, which generate a sequence of improved feasible points and converge at least to a locally optimal solution, to solve these three optimization problems. Performance of the proposed algorithms is analyzed through extensive simulations under various settings of transmit power budget, number of users, total bandwidth, transmission time, and number of transmit antennas at the BS. Simulation results clearly demonstrate the merits of the proposed algorithms.

I. Introduction

Providing ultra reliable and low-latency communication (URLLC) is a considerable challenge for 5G wireless networks and beyond [1], [2]. Reliability is typically measured by the probability that a certain finite (and usually small) amount of data is successfully transferred within a certain time duration, which requires a very low block error rate in the range of for ultra reliable communication [3]. Latency, on the other hand, is governed by the communication delay, which includes processing delay, over-the-air transmission delay, and any delay encountered due to packet retransmissions, if required. At present, low latency services are targeting even less than a millisecond. This makes the design and provision of URLLC challenging because low-latency and ultrahigh reliability are two conflicting requirements [4]. Low latency requirements mandate short-packet communication, which in turn severely limits the channel coding gain or the reliability. On the other hand, securing reliability demands for more resources (e.g., parity, redundancy, and retransmissions), which in turn increases latency.