Pairwise Transmission Using Superposition Coding for Relay-Assisted Downlink Communications

We consider the downlink relay network, for which a pairwise transmission strategy is suggested with the users being grouped into pairs and the source communicating to each pair of users using a novel pairwise relaying protocol. By utilizing superposition coding (SC), the messages of each two paired users are intentionally transmitted non-orthogonally or simultaneously in the same time-frequency channel to compensate for the bandwidth loss of relaying. We investigate the optimal decoding at the relay and the users, respectively, which are then combined to guide the selection of the power allocation factors of SC and to decide the corresponding decoding order. Then, a largely simplified outage expression of the pairwise relaying protocol is given, based on which we investigate, for the multi-user scenarios, the user pairing problem from the perspective of minimizing the total power consumption under targeted outage probabilities of the users. An easily implemented user pairing algorithm with computational complexity of O(K-2) is provided, which achieves a performance that is very close to that of the optimal user pairing which can be solved in polynomial time, O(K-3). Numerical results show that the pairwise transmission strategy can significantly decrease the average transmit power.