|dc.description.abstract||The sensor nodes in Internet-of-Things (IoT) systems are severely constrained by the amount of battery energy, which limits the network lifetime and thus the quality of service. Recently, the radio frequency (RF) energy harvesting (EH) technique has been proposed as a new viable solution to prolong battery longevity in order to supply continuous and stable energy to wireless devices over the air. To this end, a new research paradigm, termed wireless-powered communication network (WPCN), was proposed and inspired many new research branches. In a WPCN, wireless devices have no embedded energy supply and only use the harvested RF energy to perform information processing/transmission. According to the current research background on WPCNs, one transmission block is divided into two time slots: EH phase and the information transmission (IT) phase. The EH node will exhaust all the energy it harvests in the EH phase to transmit data in the IT phase. This time-division (TD) based scheduling scheme may lead to sub-optimal system performance, as the amount of harvested energy is normally very limited during each scheduled slot and cannot support effective IT. In this case, the EH devices may harvest and accumulate energy for several consecutive blocks before being scheduled to perform one effective IT. Thereby, in this thesis, I consider the design of scheduling schemes and performance evaluation for wireless powered IoT systems with energy accumulation (EA).
In this thesis, I first discuss the multi-user scheduling problem in a wireless powered IoT system with EA. Although multi-user scheduling has been extensively investigated in many published studies, the existing multi-user scheduling schemes are designed for conventional wireless communication networks or the aforementioned TD-based WPCN system. There are few studies of user-scheduling problems in WPCN with EA. On the other hand, to further improve the spectral efficiency, I introduce the full-duplex (FD) technique as it enables wireless devices to transmit and receive data simultaneously in the same frequency band while the receiving antenna will suffer from self-interference (SI). To this end, in this thesis, I start with a multi-user FD wireless-powered IoT system consisting of a FD hybrid access point (HAP) and multiple half-duplex (HD) IoT devices (IoDs). All IoDs are equipped with one antenna, while the HAP is equipped with two antennas. The HAP's two-antenna structure enables its FD working mode, in which one antenna is used to receive the scheduled IoD's signal in the uplink (UL) and the other is used to wirelessly charge the remaining IoDs by broadcasting RF signals in the downlink (DL). All IoDs have no embedded energy supply and thus need to perform EH before transmitting their data to the HAP. To maximize the system average throughput, I first design a new throughput-oriented scheduling scheme. Secondly, to strike a balance between the system throughput and user fairness, I then propose a fairness-oriented scheduling scheme. The system outage probability and average throughput are analyzed. Simulation results validate the performance analysis and demonstrate the performance superiority of both proposed schemes over existing schemes.
I then consider the scheduling schemes in security solutions for the physical layer of the wireless-powered IoT system. Besides the energy limitation, another major factor which acts as a limitation for IoT is security. In a traditional IoT communication network with eavesdropper, to improve the network security using solutions on the physical layer, one or more friendly jammers, either with multiple antennas or with a single antenna, transmit controllable artificial noises (ANs) to cooperatively resist eavesdropping. With deliberate design of ANs, such cooperation between the source node (SN) and jammers can confound eavesdroppers without yielding any interference at the legitimate receiver. However, despite its benefits, a practical challenge of this cooperative jamming (CJ) is its detrimental effect on the lifetime of the wireless network. Therefore, wireless EH techniques that enable the jammers to harvest energy from the ambient RF signals make practical sense in this context. Moreover, how to schedule different EH jammer(s) to jam the eavesdropper is also an open question worthy of investigation. This is because, in a multi-jammer scenario, if too many jammers perform CJ in the current block, in upcoming blocks, there may be few jammers with sufficient energy to perform effective CJ. Inversely, if most of the jammers always keep harvesting and accumulating energy and do not schedule to perform CJ regularly, the system could suffer from unnecessary secrecy outage. To this end, in this thesis, I propose two jammer scheduling schemes in a wireless-powered IoT system consisting of multiple EH friendly jammers and one eavesdropper. One is accumulate-then-jam (ATJ) with single jammer selection (SJS) and the other one is ATJ with multiple jammer selection (ATJ-MJS). The ATJ-SJS scheme allows, at most, one jammer to perform CJ during one transmission block. ATJ-MJS allows more than one jammer to jam the eavesdropper per block. Simulation results validate the performance analysis and demonstrate that the ATJ-SJS is superior to the ATJ-MJS if the number of jammers is small and vice versa.
Finally, I discuss the scheduling of the data packet for wireless-powered IoT systems in a finite blocklength (FBL) regime. Recently, research results on the capacity analysis of the IoT system shows that the Shannon capacity, with the ideal assumption of communicating with large enough blocks, is not necessarily appropriate to approximate the channel capacity of both conventional and wireless-powered IoT systems. This is because the wireless-powered IoT system: 1) mainly focuses on low power applications; 2) requires strict reliability of the energy supply and of the data transfer; and 3) conveys the information in short-packets with FBL. The third requirement is caused by the fact that most IoT applications only require transmission of small data size, which means the blocklength cannot be treated as arbitrary long. When the blocklength is short, the error probability becomes significant even if the selected rate is below the Shannon limit. The maximum achievable rate in this case should be approximated by jointly considering the blocklength and the allowed error probability. To this end, as an initial effort on the study of FBL based wireless-powered IoT system, in this thesis, I consider the packet scheduling problem through a simple end-to-end communication with FBL assumption. The system consists of one EH SN and one HAP. The SN has no constant power supply and has to harvest and accumulate from the signals broadcast by the HAP before transmitting information. I use the partially observable Markov decision process (POMDP) to formulate the system and find the optimal packet scheduling policy through dynamic programming to maximize the long-term throughput, given the estimated link quality and the current residual energy. The numerical results show that the POMDP based method presents higher long-term throughput than the myopic method, as the proposed POMDP based method considers the energy causality.||en_AU|
|dc.publisher||University of Sydney||en_AU|
|dc.publisher||Faculty of Engineering and IT||en_AU|
|dc.publisher||Department of Electrical and Information Engineering||en_AU|
|dc.rights||The author retains copyright of this thesis. It may only be used for the purposes of research and study. It must not be used for any other purposes and may not be transmitted or shared with others without prior permission.||en_AU|
|dc.subject||Scheduling scheme design||en_AU|
|dc.subject.other||POST DG EXPORT SUBMISSION||en_AU|
|dc.title||Scheduling Schemes in Wireless Powered IoT Systems: Protocol Design and Performance Evaluation.||en_AU|
|dc.type.pubtype||Doctor of Philosophy Ph.D.||en_AU|
|dc.description.disclaimer||Access is restricted to staff and students of the University of Sydney . UniKey credentials are required. Non university access may be obtained by visiting the University of Sydney Library.||en_AU|
|Appears in Collections:||Sydney Digital Theses (University of Sydney Access only)|