Energy Efficient Resource Allocation for Wireless Powered Communication Networks

The exponential growth of smart wireless devices has put much pressure on the spectral efficiency and energy efficiency (EE) of the Internet of things (IoT) networks and wireless sensor networks. In order to support energy constrained wireless devices, wireless powered communication networks (WPCN) have been proposed based on different wireless powered transmission (WPT) technologies, e.g., simultaneous wireless information and power transfer (SWIPT), harvest then transmit (HTT) and backscatter communication (BackCom). We note that energy-efficient resource allocation schemes need to be tailored to the different WPT technologies used in WPCNs. In this thesis (including four papers), we classify WPCNs into three types according to the way of information transmission: active transmission, passive transmission and hybrid transmission, and present energy-efficient resource allocation schemes for them in different scenarios of WPCNs.

In active transmission-based WPCNs, a radio frequency (RF) power source, e.g., a base station (BS) or a power beacon (PB), sends an RF signal to a transmitter, which harvests energy from the received RF signal through its energy harvesting (EH) circuit and generates its own RF signal to carry information to a receiver. In Paper I, we consider a SWIPT-enabled device-to-device (D2D) underlaid network, where a D2D receiver decodes information and harvests energy from its associated D2D transmitter simultaneously via its SWIPT circuit, and propose to maximize the sum EE of all D2D links by optimizing the spectrum resource and power allocation, and the power splitting ratio of each D2D device based on a non-linear EH model. We find that the number of SWIPT-enabled D2D links that maximize the sum EE is limited by the EH circuit sensitivity, especially when the D2D communication distance is long.

In passive transmission-based WPCNs, an RF power source sends an RF signal to a backscatter device (BD), which backscatters parts of the incident RF signal to a receiver and harvests energy from the rest of the incident RF signal to support the backscatter circuit. In Paper II, we propose to ensure the max-min EE fairness among the backscatter links by jointly optimizing the PB transmission power and the backscatter reflection coefficients. Our results show that the proposed max-min EE resource allocation scheme is more effective when the throughput requirement of the BDs is lower and the channel power gain difference among different PB-to-BD links is smaller. In Paper III, we propose to maximize the system EE of a symbiotic radio (SR) network that contains a primary link and multiple BDs, each being able to harvest energy while backscattering, by optimizing the primary transmitter (PT) transmission power, the BDs' reflection coefficients and time division multiple access (TDMA) time slot durations for both the parasitic SR (PSR) and commensal SR (CSR) cases. The simulation results show that the system EE is maximized when all BDs only achieve the minimum throughput requirement in the PSR case, while in the CSR case, the system EE is maximized when a best BD that can contribute the most toward the system EE is allocated the maximum allowed time to backscatter its information to the primary receiver (PR), and this best BD is determined by the optimized PT transmission power in the corresponding time slot.

In hybrid transmission-based WPCNs, the wireless devices are equipped with both the RF signal generation circuit and the backscatter circuit to support active transmission and passive transmission, respectively. In paper IV, we maximize the total EE of all the IoT nodes, which are powered by an unmanned aerial vehicle (UAV) and need to send information to a reader, by optimizing the UAV's transmit power and trajectory, the IoT nodes' backscatter reflection coefficients, transmit power for active transmission, and time allocation between backscattering and active transmission. Our results show that the UAV tends to fly toward the IoT nodes with better channel conditions to the reader, and the maximum total EE of the IoT nodes is achieved when the IoT node that is closest to the reader achieves the highest throughput, while other IoT nodes maintaining the minimum throughout requirement.