Development and design of applications for UAV-based satellite communication terminal antenna evaluation using deep-reinforcement learning

In recent years, satellites are launched almost on a daily basis and most of them are to be operated in Non-Geostationary Orbit (NGSO). The number of user terminals communicating with satellites is rapidly increasing. The interference has become a serious issue due to the crowded communication environment and the increased popularity of NGSO. Utilization of NGSO adds more complexity for operating user terminals since it requires tracking the satellite which is not static from the terminals’ points of view. Also, the risk of interference has escalated due to the greater demand for Satellite-communication-On-The-Move (SOTM), which involves the need to keep terminals constantly pointing toward the target satellite while they are installed on a moving object. To ensure a safe communication environment, the terminal antenna must be verified based on set requirements. However, the test process at conventional test facilities is inefficient and does not have a solution to test antennas in the new communication scenarios. Therefore, this thesis aims to develop in-situ Unmanned Aerial Vehicle (UAV) -based measurement applications that are autonomously guided to enhance the efficiency of the measurement and to propose novel measurement methods to verify the antennas operated in new environments. Utilizing UAVs and performing measurements onsite is challenging due to the additional error sources and uncertainties in measurements and sensor positioning. In this work, a new deep-reinforcement learning algorithm is developed which can adapt to the dynamic environment under the presence of disturbances. Using this algorithm, the applications to verify the boresight angle offset of terminal antennas and to evaluate SOTM terminal antennas are proposed. The proposed applications are tested based on the numerical simulations and the results showed that the developed applications improved the efficiency of measurements and satisfied the required measurement accuracy. The thesis investigates novel measurement approaches for a new generation of satellite communication aiming to respond to the measurement demands that currently have no solution.