Multivariable Receding Horizon Control of Aircraft with Actuator Constraints

This paper develops a constrained Model Predictive Control (MPC) formulation for longitudinal control of a fixed wing aircraft. In order to account for the inherent coupling between the inputs and outputs, which is a characteristic of multivariable systems, multiple prediction horizons are used, i.e. one for each output. Furthermore, a novel Quadratic Programming (QP) problem is derived to solve this MPC problem, via the Primal-Dual procedure. Numerical simulations using two QP algorithms demonstrate successful tracking performance of the MPC based controller.