Model predictive control of a nonlinear laboratory twin rotor aero-dynamical system

Abstract

Model predictive control (MPC) represents a powerful technique for controlling a system that has an accurate mathematical model for describing its behavior. MPC implies solving an optimization problem (OP) in order to minimize some user defined objective function (OF) subjected to constraints on the characteristic variables. This paper presents a real-time implementation of a nonlinear MPC on a twin rotor aerodynamic system (TRAS). The mathematical model of TRAS is a nonlinear one and has only four measurable states out of six. In order to benefit from state-of-the-art OP solvers, online linearization of the TRAS mathematical model is used both for transforming the OP into a convex one and also for optimal estimation of the states using and Extended Kalman Filter (EKF) approach. Detailed implementation and validation of the proposed MPC approach is offered with insightful discussions.