Robust observer-based controllers for a tilted-motors hexacopter under external perturbations

Abstract

Unmanned aerial vehicles such as multicopters have recently been used for more complicated, dangerous, and demanding applications such as payload transportation, assistance in natural disasters, or firefighting emergencies. In the context of research and technology development, these vehicles are required to fly in harsh environments, perform complex maneuvers, and track trajectories. Due to their vertical take-off and landing capabilities, these vehicles are attractive for such applications. However, they face limitations when flying in highly perturbed environments, which can deviate the vehicle or even destabilize it and lead to loss of control. In addition, the full state required for control design is not always available and may contain measurement noise due to sensors with limited accuracy or operation in denied environments. Likewise, for attitude representation, Euler angles or rotation matrices present the well-known disadvantages of gimbal lock and discontinuities. Therefore, this thesis presents the design of a unit quaternion based differentiator, that estimate the derivative of the unit quaternion to retrieve a velocity, in addition it does not use a mathematical model of the system; as well as a quaternion based extended state observer, which allows recovering unknown states and the total external disturbances acting on the rotorcraft. Thus, observer-based controllers following an adaptive sliding mode control scheme are also proposed, which ensures finite-time convergence and robustness against bounded external disturbances and model uncertainties, while reducing the chattering effect through non-overestimation of the control gain. Moreover, computer simulations are developed and tested usingMATLAB, Simulink, and Simscape Multibody environments. Ultimately, the observer-based controllers are deployed on the flight computer of a real hexacopter platform. The results demonstrate the viability of the control scheme to track desired trajectories and the feasibility of the quaternion observer to estimate unmeasurable states.