Neural Network-Based Load Swing Angle Tracking Control for Underactuated Ship-Mounted Cranes

As a typical marine engineering equipment, ship-mounted crane systems play a crucial role in meeting the demands of maritime cargo transportation and other engineering operations. However, due to their underactuated characteristics and the uncertain marine environment, particularly the disturbances caused by waves and wind leading to continuous variations in the ship's roll angle, these systems face a series of control challenges. To address these issues, this paper proposes A sliding mode control approach leveraging neural networks to achieve stable tracking of the load swing angle with respect to the ship's roll angle. Specifically, a suitable sliding mode surface is designed to effectively control both the actuated and underactuated state variables. Concurrently, a neural network is constructed, and appropriate update rules are proposed to effectively compensate for the unmatched disturbances caused by wave-induced ship motions and the uncertainties in the ship-mounted crane system. Based on the proposed control method, the trolley can quickly reach its preset position, and the load swing angle can dynamically track the ship's roll angle, achieving load swing suppression in the global coordinate system. Through rigorous analysis using Lyapunov techniques, the asymptotic stability of the predefined sliding mode surface and state variables is theoretically proven. Finally, a series of simulation results demonstrate the effectiveness and strong robustness of the proposed neural network-based sliding mode controller.