In this paper, we consider a hybrid underlay-overlay cognitive radio with energy harvesting. In the considered system, secondary user can harvest energy from the primary user's signal as well as from the other ambient sources, such as solar, wind, vibration and so on. Energy is harvested from the primary user's signal when the primary channel is found in busy state. The secondary user either operates in one of the two transmission modes; overlay and underlay in order to maximize the throughput, remains in the sleep mode in order to conserve energy, or harvests energy from the primary channel in order to maximize the remaining energy. To maximize long-term throughput of the system, we propose an access strategy in which the partially observable Markov decision process framework is used to determine action of the secondary user, and energy threshold is used to determine the transmission mode (overlay or underlay) of secondary user. Simulations show that for certain values of the system parameters, the considered system provides 60 improved throughput than overlay-only cognitive radio and 43 enhanced throughput than a hybrid cognitive radio system harvesting energy only from the ambient sources. However, increasing the throughput also increases computational burden on the secondary user, which may increase latency and energy requirements of the system.
展开▼
