Cognitive Relay Networks: Opportunistic or Uncoded Decode-and-Forward Relaying?

This paper analyzes the outage performance of cognitive relay networks operating in the same frequency band of a primary transceiver. Different from previous works, opportunistic decode-and-forward (ODF) and uncoded decode-and-forward (UDF) relaying schemes are considered, in which the relay makes either a hard decision on whether or not to transmit or generates soft information of the received signal by using the maximum-likelihood (ML) detection. To facilitate the coexistence between the cognitive user (CU) and the primary user (PU), the transmit power of both the source and the relay should be optimized to minimize the outage probability of the CU, while satisfying both the outage probability constraint on the PU and their transmit power constraints. The resulting optimization problem for the ODF scheme is generally nonconvex, but the suboptimal solution can be obtained when the decoding rate threshold of the relay is sufficiently small. Whereas for the UDF scheme, the suboptimal solution is obtained by minimizing the upper bound of its outage probability. For comparison, traditional DF relaying with optimal power allocation and direct transmission are also investigated. Simulation results confirm the tightness of suboptimal solutions and the negligible performance gap for both the ODF and the UDF schemes, and demonstrate the performance improvement provided by both schemes, particularly when the relay is far from the primary receiver.