Resource Optimization For Energy Harvesting Wireless Communication System

The rapid development of the communication network profoundly improves the quality of people's life, but also faces a severe energy consumption problem. The en-ergy harvesting (EH) is emerging as a promising technology conceived for alleviat-ing energy shortages and reducing emissions, which collects natural renewable energy from the ambience for powering the electrical equipments. Besides, the wireless en-ergy transfer technology also provides a new way for wireless remote energy supply.However, the stochastic nature and complexity of wireless network pose immense chal-lenges of applying energy harvesting technique. Thus, how to overcome the network uncertainty and effectively use the harvested energy for achieving data transmission have realistic significance, as well as improving the efficiency during process of wire-less energy transfer. By using the theory and technology of control optimization, this dissertation conducts a research on application of energy harvesting technique, includ-ing the transmission optimization in process of wireless energy transfer and adaptive resource scheduling for video transmission in EH network. The main contributions of this dissertation are summarized as follows:1) Researching on the efficient transmission of wireless energy in wireless envi-ronments. In wireless networks, the random variation of wireless channel and the uncertainty of energy request greatly influence the efficiency of wireless en-ergy transfer. High-quality channel condition is beneficial for improving trans-mission efficiency, while poor channel leads to huge energy consumption. To overcome these difficulties, we model the process of wireless energy transfer as a Markov decision process, and investigate the details in two different scenar-ios, namely non-interruptible case and interruptible case. Then, the stochastic dynamic programming is applied to derive online optimal threshold methods for non-interruptible case and interruptible case, respectively. Finally, the appropri-ate times for wireless energy transfer are achieved by the comparison between the actual transmission energy consumption and the optimal threshold. The sim-ulation results demonstrate that the proposed algorithm is capable of effectively reducing the energy consumption during the process of wireless energy transfer.2) Conceiving an adaptive policy for optimally utilizing the harvested energy to-wards meeting the specific requirements of video streaming services. Due to the stochastic nature of wireless communications and harvested energy, prompt video bitrate adaptation is highly desired for accommodating the time-varying network-load conditions. This dissertation constructs a mathematical model for scalable video adaptive transmission. Since a higher video bitrate is more likely to deplete the energy in the battery, we define the metric of energy starvation probability (E-SP) for characterizing the degree of matching between the energy demand and the harvested energy. An ESP estimation model based on Large Deviation Principle(LDP) is proposed by monitoring the energy-buffer fullness and its variations in the rechargeable battery, which can accurately characterize the probability of the rare events of energy starvation mode and assist the transmitter in achieving a fine control of the transmitted video enhancement layers. Furthermore, we in-voke classic perturbation theory for evaluating the impact of transmitting more video layers on the ESP. This method assists us in deriving an improved layer switching algorithm having a significantly lower transmission interruption rate and reduced flickers imposed on the video quality. The simulation results show that the proposed method achieves an improved performance.3) Focusing on the optimization problem of multi-user scalable video transmission in energy harvesting wireless communication network. When considering the video service for multiple users, there exist new challenges because of the complexity of network, the coordination of multiple users, and the uncertainty of harvested energy. Thus, we formulate the scalable video transmission for multiple users in EH network as a constrained stochastic optimization problem, which maximizes the system utility characterizing the perceivable quality subject to the smooth playback and energy availability constraint. Based on the Lyapunov stochastic optimization technique, an average data queue and a energy queue are construct-ed and the original long-term average utility problem is decomposed into three instantaneous optimization subproblems. By exploiting the special structure of each subproblems, we derive the energy harvesting determination, video layer selection, transmission assignment and power allocation that operates at time-scales. Finally, several experiments relying on real video traces are conducted for demonstrating the achievable performance of the proposed method. The experi-mental results illustrate that the proposed solution can provide a prompt response to the variation of the environment.