Contrôle non linéaire par backstepping d'un hélicoptère de type quadrotor pour des applications autonomes

Abstract

The quadrotor aircraft is a class of helicopter, more specifically of multirotors. The quadrotor has several characteristics (mechanically simple, vertical takeoff and landing, hovering capacities, agile) that give it several operational advantages over other types of aircraft. But its beneficts come at a cost : the quadrotor has a highly nonlinear dynamics, coupled and underactuated which makes it impossible to operate without a feedback controller action. This thesis proposes to design a nonlinear backstepping control law to control the exact position of the quadrotor in space. This controller can then be used to perform autonomous missions with a quadrotor. The application we are interested in this thesis comes from the research project « Launch and Forget : Aerial Relay Node » of the École de technologie supérieure (ÉTS) in partnership with the company Ultra-Electronics. It seeks to develop and implement an innovative autonavigation law allowing the use of a drone as a telecommunication link. The overall objective of the research project presented in the thesis is to design a backstepping controller for the navigation of a quadrotor according to predefined waypoints and provide the features necessary to implement the autonavigation law developped in the project « Launch and Forget : Aerial Relay Node ». To do this, a nonlinear model of the quadrotor is developped. A backstepping control based on this model is design and its stability is validated using Lyapunov theory and Input State Stability (ISS). An exact derivative estimator based on a second order sliding mode is used in the design of the controller to calculate the virtual control high order derivative. The controller is validated by simulation and by practical flight tests using the Pelican quadrotor manufactured by the company Ascending Technologies.

Date	20 Aug 2014
Original language	French
Awarding Institution	École de technologie supérieure
Supervisor	Ouassima Akhrif (Supervisor)