Sum-Rate Maximization Schemes for $K$-User MISO Interference Channels With a Cognitive Relay

In this paper, we consider K-user multiple-input single-output interference channels with a cognitive relay. Assuming that data of all transmitters and channel state information are known at the cognitive relay, we design a linear precoder for the cognitive relay with the aim of maximizing the sum-rate without changing the transmitter operations at all transmitters. We first define the receiver set as a set which contains a part of the receivers, and then introduce a performance metric called “partial signal-to-interference-plus-noise ratio” (PSINR) based on the receiver set. Then, we can obtain a precoder at the cognitive relay by solving the PSINR maximization problem. The optimal receiver set which yields the maximum sum-rate can be identified by checking all possible receiver sets. Since this exhaustive search has prohibitive complexity, we develop a low complexity set search method by utilizing the properties of the optimal receiver set. Combining the PSINR maximization problem and the low complexity search method, we finally propose a precoder design scheme for the sum-rate maximization. Numerical simulation results confirm that the proposed scheme shows performance close to the projected gradient method with much reduced complexity.