A Joint Scheduling and Cell Clustering Scheme for MU-MIMO Downlink with Limited Coordination

In next-generation cellular systems, Multi-User MIMO (MU-MIMO) schemes are expected to achieve very high downlink throughput, provided that issues such as Inter-Cell Interference (ICI), user scheduling, etc. are properly handled. In order to accomplish this goal, we consider a joint scheduling and cell clustering scheme where groups of adjacent cells (clusters) are jointly controlled and effectively act as a distributed multi-antenna transmitter. An "overlapped" clustering scheme is used whereby different clustering patterns operate on different frequency subchannels. The arrangement is such that each user in the system is located away from the cluster-edge and sees a favorable ICI level on one or more subchannels. This leads to significant per-user and per-cell performance improvements over single-arrangement clustering schemes that force a user to use only one cluster arrangement. Critical to realizing such benefits in an overlapped cluster arrangement is the joint operation of downlink opportunistic scheduling across all cluster patterns, i.e., across the frequency bands. Indeed, joint scheduling enables users to optimally share and balance the use multiple subchannels (i.e, multiple clustering patterns). Such a balance does not trivially schedule users close to cluster centers. We demonstrate, under the popular Proportionally Fair Scheduling (PFS) criterion, that the joint system operation of scheduling and overlapped arrangements yields significant improvements to "cell-edge" user rates while also increasing overall per-cell throughput.