Abstract
As a new standard traffic service for time-sensitive applications over beyond-5G (5G+) mobile wireless networks, ultra-reliable and low-latency communications (URLLC) have received tremendous research attention. The major design issue raised by URLLC is how to support latency-sensitive multimedia transmissions while guaranteeing the reliability bound over timevarying wireless channels. Along this direction, researchers have developed a wide spectrum of advanced techniques, including statistical delay-bounded quality-of-service (QoS) provisioning, cell-free (CF) massive multiple-input multiple-output (m-MIMO), finite blocklength coding (FBC), hybrid automatic repeat request (HARQ) protocol, etc. CF m-MIMO systems, where a large number of distributed access points (APs) jointly serve all users in a coordinated manner using only local channel state information (CSI), have emerged as one of 5G+ key techniques to significantly improve various mobile QoS performances, including upper-bounded delay and error-rate, throughput, user-coverage probability, etc. However, when being integrated with millimeter wave (mmWave), how to efficiently characterize the networking dynamics for CF m-MIMO schemes using FBC based statistical delay bounded QoS theory has imposed many new challenges not encountered before. To overcome these problems, in this paper we propose statistical delay/error-rate bounded QoS provisioning architectures/schemes over mmWave user-centric CF m-MIMO and FBC-HARQ based 5G+ wireless mobile networks. In particular, we establish the comprehensive mmWave user-centric CF m-MIMO based system models. We also apply HARQ-IR protocol to determine the channel capacity and characterize the QoS metrics in terms of error probability. We also conduct a set of simulations to validate and evaluate our proposed FBC-HARQ based mmWave user-centric CF m-MIMO schemes.
Funding Information
  • U.S. National Science Foundation (CCF-2008975, ECCS-1408601, CNS-1205726)
  • U.S. Air Force (FA9453-15-C-0423)
  • U.S. National Science Foundation (CCF-0939370, CCF-1513915)

This publication has 42 references indexed in Scilit: