An intelligent control method for autonomous ferries in inland waterways: A nonlinear terminal-free model predictive control approach

Ferry is an important scenario in inland waterways, and autonomy is the trend. Autonomous ferry (AF) is generally converted into the intelligent control problem of the reference path. Due to space constraints of inland waterways, network communication status, and obstacles in inland waterways, the reference path of autonomous ferry has various curves and discontinuity, which brings difficulties to the intelligent control of autonomous ferry. Therefore, a new intelligent control method is proposed based on nonlinear model predictive control (NMPC). The NMPC strategy without terminal cost and terminal set constraint is designed for the discontinuous paths, which reduces the computation burden and increases the controller's feasible domain and real-time performance. On this basis, dynamic trajectory planning and control are combined to realize the quadratic planning of the trajectory of the various curves, and artificial reference decision variables are also integrated to solve it. A method to calculate the lower bound of the prediction horizon is proposed to prove the stability of the controller and a controlled forward invariance set is also designed to establish the recursive feasibility. Moreover, virtual simulation experiments and field lake tests are carried out to verify the robustness and accuracy of the controller.