Timely Scheduling of URLLC Packets Using Precoder Compatibility Estimates

Abstract

5G systems need to efficiently support both enhanced mobile broadband traffic (eMBB) and ultrareliable, low-latency communications (URLLC) traffic. In this paper, the problem of joint eMBB and URLLC scheduling is studied with the objective of supporting the URLLC requirements with minimal impact on eMBB performance. We propose a preemptive scheduler exploiting a precoder compatibility estimate, which measures the degree of similarity between different multiple-input multiple-output (MIMO) channels. This metric allows us to determine which eMBB transmission to pause for serving a URLLC user without recomputing the precoding matrix, thus reducing the processing time needed for multiuser MIMO precoder computation, while both relaxing cell pilot periodicity and reducing URLLC demodulation pilot overhead. This is achieved thanks to the fact that precoder compatibility metrics can be acquired at the user side in an opportunistic manner based on pilot symbols destined to ongoing downlink transmissions. Simulation results assess the performance gains of the proposed scheduler in terms of both URLLC block error rate and eMBB throughput.