Model Predictive and Decoupled Thrust Allocation for Overactuated Inland Surface Vessels

This work presents and implements a decoupled Thrust Allocation (TA) algorithm for overactuated inland surface vessels. The utilised control strategy is Nonlinear Model Predictive Control (NMPC), which solves an Optimal Control Problem (OCP) over a control horizon. The controller optimises for minimum energy consumption, subject to the desired responsiveness. We describe the established NMPC formulation as a Nonlinear Programming Problem (NLP) employing the multiple shooting method. Here we use a variety of expressions for nonlinear constraints and saturation models of the propulsion system. Finally, we detail the implementation of the NLP. The resulting implementation has been evaluated on computational load and convergence, embedded on a scale model, during an on-land stationary test.