Performance of Adaptive Link Selection with Buffer-Aided Relays in Underlay Cognitive Networks

Abstract: In this paper, we investigate the performance of a three-node dual-hop cognitive radio network (CRN) with a half-duplex (HD) decode-and-forward (DF) buffer-aided relay. We derive expressions for the average rate and symbol error rate (SER) performance of an adaptive link selection based channel-aware buffer-aided relay (CABR) scheme that imposes peak-power and peak-interference constraints on the secondary nodes, and compare them with those of conventional non-buffer-aided relay (CNBR) and conventional buffer-aided relay (CBR) schemes for a delay-tolerant system. For finite-delay systems, we analyze the performance of a modified threshold-based scheme for fixed-rate transmission, and demonstrate that use of a last-in-first-out buffer is advantageous in some situations. We bring out the trade-offs between delay, throughput and SER. Computer simulation results are presented to demonstrate accuracy of the derived expressions.