Model predictive control of drone systems via stable embedding technique

Abstract

Model Predictive Control (MPC) is widely used due to its feature that it can build a desired controller by formulating optimization problems. The process of computing a MPC control law includes discretization of the system dynamics. In this procedure, if the system is defined on a manifold that is not homeomorphic to the Euclidean space, the computed trajectories may not lie on the manifold where the system is defined. In this paper, we apply MPC to drone systems with stable embedding technique that allows the design of controllers for systems defined on manifolds in the ambient spaces. Simulation results show that the proposed MPC technique overcomes the issue and outperforms the existing methods in time efficiency and state errors. Finally, the proposed algorithms are implemented into a real drone system to show that simulation results are consistent with experiments.