On optimal offline time sharing policy for energy harvesting underlay cognitive radio

Abstract

Abstract-RF energy harvesting can be used to power communication devices so that perpetual operation of such devices can be ensured. We consider a RF energy harvesting underlay cognitive radio system operating in slotted fashion. The primary user (PU) is equipped with a reliable power source and transmits with a constant power in all the slots. However, the secondary user (SU) harvests energy from primary's transmission and simultaneously transmits it's own data such that interference at the primary receiver (PR) remains below an acceptable threshold. At the secondary transmitter (ST), each time slot is divided into two phases: energy harvesting (EH) phase and information transfer (IT) phase. We formulated the problem of maximizing the achievable secondary sum rate under primary receiver's protection criteria as a convex optimization problem and obtained the optimal time sharing between EH and IT phase, and optimal secondary transmit power under offline setting. The optimal offline scheme is then compared with an online myopic policy, where the optimal time sharing between EH and IT phase, and optimal secondary transmit power are obtained based on instantaneous channel gains only.

Harvest-or-Transmit Policy for Cognitive Radio Networks: A Learning Theoretic Approach

“…In [7], the authors designed a robust transceiver for the simultaneous wireless information and power transfer (SWIPT) based multiple-input multiple-output (MIMO) underlay CRN. The work in [8], [9], and [10] considered an underlay EH-CRN where the ST harvested energy from PT's interference in [8] and [9], and from other environmental sources in [10]. The authors obtained a myopic transmission policy in [8], an optimal offline policy in [9] and a robust online policy in [10].…”

“…The work in [8], [9], and [10] considered an underlay EH-CRN where the ST harvested energy from PT's interference in [8] and [9], and from other environmental sources in [10]. The authors obtained a myopic transmission policy in [8], an optimal offline policy in [9] and a robust online policy in [10]. In [11], [12], and [13], authors considered a multihop underlay EH-CRN network where the energy harvesting ST communicated with the secondary receiver (SR) over multiple hops using the time division multiple access (TDMA) protocol.…”

“…The aforementioned works on optimal transmission policies in underlay EH-CRNs focused mainly on myopic [6]- [8], [11]- [13], offline [5], [6], [9] and online policies [10], [14], [16]. The myopic policies aim to maximize the immediate throughput and do not care about the past and future channel conditions and energy arrivals, which is suboptimal in multislot scenarios.…”

“…The offline policies, on the other hand, assume that the channel conditions and energy arrivals are known non-causally at the transmitters. This information allows the transmitters to adapt their transmission powers according to channel conditions and energy availability, which results in higher achievable throughput [9]. However, the non-causal knowledge of channel gains and energy arrivals at the transmitters may be an unreasonable assumption if the environment is highly dynamic.…”

“…The transmit power of the PT is assumed constant in all slots. Different from the system model proposed in [6], [9], [10] where each slot was shared between energy harvesting and data transmission (time-sharing model), in our model, each slot is either used for energy harvesting or data transmission. Considering harvest-or-transmit model over the time-sharing model reduces the size of action space which reduces the complexity of dynamic programming algorithm for the online policy [31].…”