Autonomous path tracking control design for a comercial quadcopter

Abstract

This paper describes the methodology followed to design a functional autopilot for the quadcopter AR.Drone developed by Parrot. The main goal is to design a control strategy for autonomous path tracking in the XY plane, comparing two different control techniques. Three phases are carried out to achieve this objective: model identification, control design and evaluation, and implementation. At the identification phase two Hammerstein models are obtained, they are characterized by having a static non-linearity preceding a linear transfer function. The control design phase is based on the use of a cascade control to regulate the position with an inner speed control loop. Furthermore, two different techniques to reduce the path tracking error are compared.