Robust global observer position-yaw control based on ellipsoid method for quadrotors

Abstract

An observed-based control strategy for position-yaw tracking of quadrotors is proposed. From a virtual controller for the position dynamics, attitude and angular velocity references are estimated with the aim of allowing the position-control components to enforce the tracking of a position reference. In addition, no measurement of the linear velocity is assumed, and the measurement of the position signal is noisy, thus, a dynamic observer for the position dynamics is employed. The ellipsoid method is considered to estimate the feedback controller and observer gains, based on an optimization procedure, in order to produce the ultimately boundedness of attitude and position tracking errors inside a compact convex vicinity of the origin, whose Lebesgue measure is minimized via LMIs. The advantages of the proposed scheme are illustrated through numerical simulations that include aerodynamic disturbances and noisy measurements. Finally, an experimental assessment is carried out to illustrate the feasibility and performance of the proposed controller in a real-time application.