User Grouping and Reflective Beamforming for IRS-Aided URLLC

This letter studies an intelligent reflecting surface (IRS)-aided downlink ultra-reliable and low-latency communication (URLLC) system, in which IRS is dedicatedly deployed to assist a base station (BS) send individual short-packet messages multiple users. To enhance the URLLC performance, users are divided into different groups for each group encoded single codeword. By considering time division access (TDMA) protocol among groups, our objective minimize total latency all subject their reliability requirements, via jointly optimizing user grouping blocklength allocation at BS together with reflective beamforming IRS. We solve minimization problem alternating optimization, blocklengths optimized by using techniques of successive convex approximation (SCA) semi-definite relaxation (SDR), while updated K-means greedy-based methods. Numerical results show that proposed designs can significantly reduce latency, as compared various benchmark schemes, unveil importance optimization exploiting joint encoding gain enhancing worst-case performance.