A purely model predictive control for a marginally stable system

Effective control of load swing in shipboard crane has been the subject of interest to both commercial and military applications. The difficulty of the control problem due to uncertainty in the nonlinear models of the crane, and the presence of numerous disturbance sources in the environment is well documented. Model predictive control (MPC) has been successfully applied in many real-world industrial applications due to its inherent robustness to modeling errors and the ability to explicitly include constraints in the problem formulation. This paper proposes a formulation that enables the use of a purely MPC-based technology for the control of load swing in shipboard cranes in real-time. We show that a decomposition of the crane model into linear state dynamics, and nonlinear static output mapping enables the use of the MPC approach in a computationally efficient way. We use our newly developed nonlinear solver to enhance the computational properties of the MPC-based approach. Simulation results are presented to demonstrate the feasibility of the proposed approach. To the best of our knowledge, this is the first successful implementation of a purely MPC-based controller for the load swing in a shipboard crane. The solution presented in this paper is ideally suited for indirect adaptive control and hence suggests an exciting opportunity for a versatile controller in a difficult application.