The design of control laws for miniature rotorcraft-based Unmanned Aerial Vehicles (UAVs) is complex due to the nonlinearity of their dynamics and their high sentitivity to aerologic perturbations. In this thesis, two classes of controllers are investigated : model predictive controllers and partial state feedback controllers when the velocities of the vehicle are not measured. Some model predictive controllers have already been developed for miniature UAVs. However, control strategies with guaranteed stability have not been simultaneously considered with models which are representative enough of the dynamics of a miniature Vertical Take Off and Landing (VTOL) vehicle. In this study, several model predictive control algorithms are proposed as well as their application to the design of guidance and control laws for VTOL UAVs. Six degrees of freedom nonlinear models are considered and stability of the closed loop system is analyzed. Stabilisation around a fixed point, trajectory tracking as well as obstacle avoidance are addressed. In some experimental or practical uses of a miniature UAV, velocity measurements may not be available. Partial state feedback or observer-based control strategies can hence be applied. To realise a good trade off between these two methods, and simultaneously obtain good performances for the closed loop system and decrease the complexity of the approach (computation delays, stability analysis), we propose a partial state feedback strategy based on the intoduction of virtual states in the system dynamics and on Lyapunov analysis. Our contribution also concerns stability analysis of a hierarchical controller, using singular perturbation theory.
