Statistical Performance Evaluation for Energy Harvesting Communications based on Large Deviation Theorem

Energy harvesting (EH) is a promising technology for enhancing a network’s quality of service (QoS). EH-based communication systems are studied by tackling the challenges of energy-outage probability and energy conditioning. These issues motivate this research to develop new solutions for increasing the lifetime of device batteries by leveraging renewable energy sources available in the surrounding environment, for instance, from solar and radio-frequency (RF) energy through harvesting. This dissertation studies an energy outage problem and user QoS requirements for energy harvesting communications. In the first part of this dissertation, the performance of an energy harvesting communication link is analysed by allowing a certain level of energy-outage. In EH systems, energy consumed from the battery depends on the QoS required by the end user and on the channel state information. At the same time, the energy arrival to the battery depends on the strength of the power source, solar in this case, and is independent of the fading channel conditions and the required QoS. Due to the independence between the energy arrival into the battery and the energy consumed from there, it is challenging to estimate the exact status of the available energy in the battery. An energy outage is experienced when there is no further energy for the system to utilise for data transmission. In this part, a thorough study was carried out to analyse the required energy harvesting (EH) rate for satisfying the QoS requirements when a level of energy-outage is allowed in a point-to-point EH-based communication system equipped with a finite-sized battery. Furthermore, an expression relating the rate of the incoming energy with the fading channel conditions and the minimum required QoS of the system was provided to analyse the performance of the EH-based communication system under energy constraints. Finally, numerical results confirm the proposed mechanism’s analytical findings and correctness. In the second part of this dissertation, the performance of point-to-point communications is investigated in which the source node can harvest and store energy from RF signals and then use the harvested energy to communicate with its end destination. The continuous availability of RF energy has proved advantageous as a wireless power source to support low-power devices, making RF-based energy harvesting an alternative and viable solution for powering next-generation wireless networks, particularly for Internet-of-Things (IoT) applications. Specifically, the point-to-point RF-based energy-harvesting communication is considered, where the transmitter, which can be an IoT sensor, implements a time-switching protocol between the energy harvesting and the information transfer, and we focus on analysing the system performance while aiming to guarantee the required QoS of the end user subject to system constraint energy outage. The time-switching circuit at the source node allows the latter to switch between harvesting energy from a distant RF energy source and transmitting data to its target destination using the scavenged energy. Using a duality principle between the physical energy queue and a proposed virtual energy queue and assuming that a certain level of energy outage can be tolerated in the communication process, the system performance was evaluated with a novel analytical framework that leverages tools for the large deviation principle. In the third and last part of this dissertation, an empirical study of the RF-EH model is presented for ensuring the QoS constraints during an energy-outage for Simultaneous Wireless Information and Power Transfer (SWIPT) network. We consider a relay network over a Rayleigh fading channel where the relay lacks a permanent power source. Thus, we obtain energy from wireless energy harvesting (EH) of the source’s signals to maintain operation. This process is performed using a time-switching protocol at the relay for enhancing the quality of service (QoS) in SWIPT networks. A numerical approach is incorporated to evaluate the performance of the proposed RF-EH model in terms of different evaluation parameters such as time-switching protocol, transmit power and outage. The assumptions of the large deviation principle are satisfied using a proposed virtual energy queuing model, which is then used for the performance analysis. We established a closed-form expression for the system’s probability of experiencing an energy outage and the energy consumed by the relay battery.