Integration of unmanned aerial vehicles (UAVs) as wireless relays in traditional telecommunication infrastructures has proved to be effective in mitigating network congestion, enhancing Internet services, and establishing ad-hoc networks in complicated environments. Efficient and scalable coordination mechanisms are crucial for managing and maintaining optimal, stable UAV formations, and adapting to the rapidly changing network demands. In this thesis, two fully-distributed control architectures are proposed that are based on Feedback Equilibrium Seeking, a unifying framework for designing algorithmic-based output feedback controllers. In particular, algorithmic trajectory planners are proposed that dynamically steer UAV swarms near efficient network configurations that give the best quality of service (QoS) to a group of Internet subscribers. Robustness to changes in the Internet demand for the discrete-time planner in closed-loop with a pre-stabilized UAV network is studied in a sampled-data setting. We provide a formal robust stability analysis of the obtained sampled-data system to ensure tracking of an optimal (time-varying, a-priori unknown) swarm configuration and experimentally validate the theoretical findings using real-world quadcopters. Precise convergence of a fleet of Bitcraze Crazyflies to the desired operating conditions is achieved in the presence of exogenous disturbances affecting the users’ request for Internet services, thus showcasing effectiveness of the adopted framework in modelling relevant objectives for mobile wireless networks.
Feedback Equilibrium Seeking control: https://arxiv.org/abs/2210.12088
C++ code: https://gitlab.ethz.ch/dfall/dfall-system