Improved Round-Robin User Scheduling for Two-Hop ZF MIMO Relay Systems

In this paper, we propose an improved round-robin scheduling (RRS) method for multi-pair two-hop MIMO relay systems, where the relay is equipped with large-scale antenna arrays and performs zero-forcing processing. Our method not only preserves the advantages of RRS, i.e., the instantaneous channel state information independence and the extremely-low computational complexity, but also achieves much higher sum rate than the conventional RRS. First, we derive a closed-form approximate expression of the asymptotic sum rate. Then, the optimal number of simultaneously transmitted data streams, K*, is obtained by an offline one-dimensional search to maximize the approximate asymptotic sum rate. Finally, we modify the traditional RRS method by setting the number of active data streams at each time slot to K*. Numerical results show that compared with the common RRS method, the proposed method provides significant sum rate improvements without any online complexity increases.