An Efficient System Orchestrator and a Novel Internet of Things Platform for Unmanned Aerial Systems

Abstract

Unmanned aerial vehicles (UAVs) are used to provide diverse civilian, commercial, and governmental services. In addition to their original tasks, UAVs can also be used to offer numerous value-added internet of things services (VAIoTSs). Sharing UAV infrastructure to provide IoT services can lower both capital and operational expenses to create a novel ecosystem with new stakeholders. The deployment of UAV-based IoT platforms, along with mechanisms to provide VAIoTSs from the sky, comes with a number of challenges, and solving few of them is the objective of this dissertation. To enable VAIoTS delivery, this dissertation introduces an innovative UAV-based IoT platform and envisions a UAV-based communication architecture including a system orchestrator (SO). The SO manages the UAVs and their on board IoT devices and optimally carries out diverse IoT services in an energy-efficient manner, using reliable data communication. Furthermore, to ensure efficient delivery of VAIoTSs using UAVs, a UAV selection mechanism is proposed, which takes into account multiple metrics in the UAV selection process such as UAVs' onboard IoT devices, energy, geographical proximities, and the priority levels of the IoT events. In this light, the following three complementary solutions are proposed: energy-aware UAV selection (EAUS), which aims to reduce UAVs' energy consumption; delay-aware UAV selection (DAUS), which aims to reduce UAVs' operational time; and fair trade-off UAV selection (FTUS), which ensures a fair trade-off between UAV energy consumption and operation time. The results demonstrate the efficiency and robustness of the proposed schemes. In addition, to provide reliable communication for the UAVs, a connection steering mechanism between mobile networks and UAVs is presented. This mechanism selects the network with the strongest radio signal among available networks in order to ensure network coverage and increase transmission rates. The efficiency of this mechanism is examined using two different LTE-4G networks of two different network providers. The results prove the efficiency of the proposed mechanism in terms of an increase in data transmission rate and a reduction in UAV energy consumption. Furthermore, as an envisioned application of the UAV-based IoT platform, a crowd surveillance use case based on face recognition is proposed. To evaluate the use case, the offloading of video data processing to a mobile edge computing (MEC) node is compared to the local processing of video data on board a UAV. The results prove the efficiency of the MEC-based offloading approach in saving the energy of UAVs. In Conclusion, the research presented in this dissertation, through the presented UAV-based IoT platform, the system orchestrator, and the designed mechanisms enables the effective delivery of VAIoTSs from the sky.