Control of quad-rotor UAVs using switched-system synthesis methods

This thesis applies switched systems synthesis and linear quadratic regulator (LQR) theory to control of a quad-rotor unmanned aerial vehicle (UAV). The thesis presents the development of the system dynamics, the theory of LQR and its implementation, the synthesis and simulation results of switched control of the UAV, which consists of a central rigid body and four propellers in a cross configuration. Since first introduced in 1917, UAVs have been extensively studied and utilized in various circumstances that prefer no human pilots aboard, due to safety, expenses, etc. Stability is crucial in controller design, while other parameters also draw great concerns, depending on the environment. The methodologies of LQR control and semidefinite programming (SDP) are discussed to provide preliminary knowledge of the switched control. Benefits of the LQR control include tracking of reference trajectories and cost function minimization. The core of switched control methods is the design and analysis of systems whose dynamical models and performance specifications are governed by the modes of an automaton. By assigning the weights properly on the performance states, the controller allows transitions between modes with stability guaranteed. The model of the UAV was established by analyzing the equations of motions based on kinemics and dynamics, then linearized and discretized for design purposes. Both the LQR and switched controllers were generated and simulated using MATLAB, and the LQR controller was transferred to the physical UAV for test and data collection. To incorporate with reality, lags to commands and saturation of the motors were taken into consideration