With the rapid development of mobile technology, more and more devices connect to the Internet of Things (IoT). The management of such large-scale networks becomes a challenge. Firstly, a large number of heterogeneous devices are distributed over a wide area, leading to a variation of the requirements of users, the performance of mobile devices, and the application scenarios. As the size of the IoT increases, the complexity of controlling such systems becomes a challenge. Most existing solutions choose global control, and are designed for a specific type of application scenario. However, any changes in the network, e.g. topology, node density, etc., affect the control schedule of the central node. Once the context changes beyond the adaptation ability, the system can hardly function anymore. Furthermore, the center node is the single break point in the control structure. Therefore, it is critical to find a solution with autonomous management, in which networks are organized and controlled by the local management of each node. Secondly, maintaining the power supply for a large number of battery-operated mobile devices in the IoT becomes a challenge. The most direct solution is to replace batteries of devices periodically. However, this costs much money, time, and human resources. Increasing the size of the battery is another commonly used approach, but this enlarges the formand weight of devices, which is unsuitable for application scenarios where size and weight of devices should be minimized. Therefore, we need an approach where devices have autonomous energy, in which batteries of mobile devices can be wirelessly charged. Based on the motivation above, the research of this dissertation is positioned in the area of autonomic computing. The proposed systems are self-adaptive self-organized and use radio-frequency based wireless power transfer. Specifically, nodes in the network can achieve global operation, based on local information exchange and control of each node, and increase battery lifetime by harvesting energy from transmitted radio waves and decreasing the duty cycle of radio in the communication protocol. In the area of self-adaptive self-organization systems, we explore controlling networks based on local information exchange. The global operation of the whole network is controlled by local management of each node. The advantage is that nodes do not need to collect a large amount of global information, which largely decreases the communication complexity of the network. We leverage this mechanism in two case studies. First, we target data aggregation in mobile networks. Our algorithmuses evolutionary dynamics to select and spread the configuration of each node, and the network automatically adapts to the variation of application scenarios. The network can optimize configurations without predesigned setup for a specific scenario. In the second case study, we design an algorithmto achieve distance estimation with self-organization in large-scale mobile networks. The algorithm uses messages collected by local information exchange for statistical calculation, and the network collectively estimates distances between nodes in the network. This improves the accuracy and extends the application area of the existing distance estimation approaches. In the area of wireless power transfer systems, the main contribution is based on the exploration of increasing the efficiency of energy transmission and utilization in mobile devices using radio-frequency based wireless power transfer. First of all, we exploit the properties of active and backscatter radio for increasing the energy efficiency of harvesters. We demonstrate the world’s first hybrid radio platformthat combines the strengths of active radio (long range and robustness to interference) and backscatter radio (low power consumption). We design a switching mechanism that selects active radio or backscatter radio for different radio channel qualities. The measurement results onmobile devices prove that harvesting and saving radio energy is not the only choice to provide autonomous energy, and that backscatter radio for communication is more energy efficient for some applications on mobile devices. Second, we save energy on the charger side to make wireless power transfer green. Wireless power transfer based on radio frequency radiation and rectification is fairly inefficient due to power decaying with distance, antenna polarization, etc. To save energy in chargers, we monitor the idle charging state in wireless power transfer networks and switch off the energy transmitters when the received energy is too lowfor rectification. Although this systemdoes not directly increase the efficiency of the radio harvesting process, the saved energy in chargers largely boosts the energy efficiency of the whole wireless power transfer network. The system is especially valuable for increasing the lifetime ofmobile chargers powered by batteries. Finally, to demonstrate the value of energy autonomy in real applications, we select indoor localization using wireless power transfer as a case study. We design a battery-less indoor localization system that can operate perpetually under wireless power transfer. The novel localization method operates at energy levels that are within the energy budget provided by wireless power transfer today, and the communication schedule is well-designed to minimize the amount of idle listening. We use off-the-shelf devices to implement and deploy the system. It proves the feasibility of using long-range wireless power transfer for mobile systems.
|Qualification||Doctor of Philosophy|
|Award date||2 Dec 2016|
|Publication status||Published - 2016|