Hybrid Predictive Control of Quadcopters Based on GOBF Models

Abstract

This article presents an approach for the control of the dynamical response of quadcopters. This is obtained by the development of a hybrid control system composed of model predictive controllers and logical rules. For the design of the predictive control system, mathematical models used for output prediction are obtained. These models are derived using Generalized Orthonormal Basis Functions. Such models are obtained from I/O data collected from the system in different modes of operation. The logical rules defined are responsible for deciding which models will represent the system according to the mode of operation. In order to validate the control system, a controller for quadcopter altitude is developed. Since quadcopters are nonlinear systems, they present different modes of operation justifying the choice. Results demonstrate the efficiency and the validity of the proposed methodology.