Augmented model predictive control of unmanned quadrotor vehicle

Abstract

Model Predictive Control (MPC) is a type of controller which generates control input by solving an optimization problem under the control and output constraints. The optimization problem consists of a cost function in which the future states of the plant of interest are also considered. In this work, an augmented version of MPC is used to control the linear model of quadrotor UAV. In general control of quadrotor is used by PID controllers implemented separately for all the control axis, but in MPC, state-space model of the vehicle is used and future states are predicted. Based on the expanded statespace matrices the control inputs over the control horizon is found with the consideration of the whole system by considering the saturation limits of inputs and limits of outputs or states which is the very powerful aspect of the controller.