Neural-Network-Based Distributed Formation Tracking Control of Marine Vessels With Heterogeneous Hydrodynamics

In this paper, a robust scheme is presented for distributed formation tracking control of marine vessels with heterogeneous hydrodynamics. Provided that multiple ships sail in proximity at sea, the hydrodynamic forces and moments would be changed greatly due to the ship-ship interactions, thereby causing the heterogeneous hydrodynamics; nevertheless, it is usually ignored in formation control of marine vessels. The heterogeneous hydrodynamics is very difficult to model and induces the uncertainty; therefore, it is treated as unknown dynamics in this paper. The neural network (NN), which is well-known for approximation-based control, is employed to handle the unknown dynamics. Note that the NN is employed to approximate the unknown dynamics collectively for reducing the learning parameters. Furthermore, unlike traditional formation control of coordination of positions, the consensus mechanism of velocities is also included into the control design to ensure the performance of formation maintenance. The presented distributed controller for each ship only utilizes the information from itself and its neighbors, aiming to prevent single-point failure in harsh marine environment. All the closed-loop signals are proved to be stable based on Lyapunov theory. Various simulations are conducted to validate the effectiveness of proposed control design.