Throughput of an Energy Harvesting Cognitive Radio Network Based on Prediction of Primary User

Abstract

In this paper, a novel prediction based cooperative spectrum sensing scheme is investigated on the performance of an energy harvesting cognitive radio (CR) network. The spectrum sensing scheme is redesigned to protect the quality of service (QoS) of primary user (PU) and to improve the utilization of spectrum holes. The decision about the PU spectrum status and energy harvesting (RF and non-RF) of a CR node are based on prediction as well as sensing at individual CR level. We consider simultaneous spectrum sensing and energy harvesting scenario through the incorporation of an energy splitting device. A CR harvests from non-RF resources if both the decisions (decision of prediction and decision of spectrum sensing) do not match or if both the decisions match in favor of the absence of PU. On contrary, it harvests from RF resources while both the decisions match in favor of the presence of PU. A CR node transmits only if both decisions indicate the absence of PU. A CR user opportunistically uses the PU spectrum for its transmission purpose under a collision constraint. The collision constraint gives an extra protection to the QoS of PU on rearrival of PU. Novel analytical expressions for detection performance, harvested energy, and network throughput are developed. The impact of prediction and other network parameters such as number of detection frames, number of cooperative CR user, splitting parameter, collision probability on throughput performance is investigated. Improvement in spectrum reuse and energy penalty during harvesting is indicated. Impact of noise power estimation on the sensing performance is also studied.