Stratégies de commande de véhicules à déplacement vertical.

Abstract

This work addresses the study of the control of autonomous Vertical Take-Off and Landing (VTOL) aircrafts. We focus on multi-rotor Unmanned Aerial Vehicles (UAVs). We first study the dynamic modeling of this class of UAVs. Using the Euler-Newton formalism, the differential equations are derived as a generalized model to describe the behavior of a multi-rotor UAV with an even number N of rotors. In a second step, we analyze the control approaches used in the literature for multi-rotors stabilization and path / trajectory tracking. Based on the dynamic model developed, we propose a hierarchical structure for the vehicle control. The proposed control approach consists of combining the Proportional, Integral, Derivative (PID) algorithm with the integral Back-stepping algorithm. We also discuss the effect of a complete rotor fault on the behavior of the vehicle. The guidance of UAVs is another element addressed in this work. We present a study of the different algorithms applied for the planning of 3D paths. We then present our contribution in the path planning for UAVs. It is a method of path planning, minimizing the path length and maximizing the distance to obstacles, based on the skeletonization of the 3D environment while taking into account the size of the vehicle. The proposed method has been tested and proves capable of producing optimal solutions in terms of safety and path length.