Master Thesis Defense - Aclecio de Jesus dos Santos скачать в хорошем качестве

Master Thesis Defense - Aclecio de Jesus dos Santos

Master Thesis defense of the Graduate Program in Electrical Engineering at UFMG Abstract: This dissertation addresses the modeling and control of an unmanned aerial vehicle (UAV) in the Quadtilt-rotor type, an aircraft capable of performing vertical takeoff and landing (VTOL), hovering, and achieving high speeds in cruise flight. The dynamic modeling of the system is based on the Newton-Euler formulation, in which the UAV is represented as a multibody system. The Quadtilt-rotor UAV poses significant challenges for modeling and control due to its highly nonlinear dynamics and the complex interactions among the coupled bodies of the system. These interactions include gyroscopic effects, Coriolis and centripetal forces, coupling between the tilting mechanisms, and aerodynamic effects associated with the wings, canards, vertical stabilizers, and the fuselage itself, all of which are accounted for in the modeling process. To ensure a global and singularity-free description, the pose of the rigid bodies that compose the aircraft is parameterized on an extended manifold of the group SE(3). As a result, the evolution of the dynamic system takes place in the space R^3 x SO(3)^5. Moreover, as an underactuated mechanical system subject to nonholonomic constraints, stabilizing the UAV around an equilibrium point is a particularly challenging task. Systems with such characteristics cannot be globally and asymptotically stabilized at a fixed equilibrium through a continuous feedback control law. Given this limitation, the present work adopts the Model Predictive Control (MPC) technique, which is particularly suitable for handling such constraints by employing a discrete-time predictive model to optimize control actions. Based on this framework, two Nonlinear MPC (NMPC) strategies are developed. The first strategy is based on a simplified representation of the UAV that neglects aerodynamic surface effects, making it more suitable for maneuvers near hovering and at low speeds. The second strategy, in contrast, employs the full model, incorporating the aerodynamic forces acting on the wings, canards, vertical stabilizers, and fuselage, thereby enabling more accurate control throughout the entire flight envelope, including high-speed regimes and transitions between helicopter and airplane flight modes. Both strategies are formulated directly on the extended SE(3) space, eliminating the need for predefined trajectories and specific terminal conditions. The controllers continuously optimize the control actions to guide the system toward equilibrium states while ensuring compliance with constraints on inputs and states. Finally, closed-loop stability and recursive feasibility analyses are conducted, and the effectiveness of the proposed approaches is demonstrated through numerical simulations in a virtual environment.