Adaptive Hierarchical Sliding Mode Control for Double-Pendulum Gantry Crane Based on Neural Network

Gantry cranes play a pivotal role in industrial production. Gantry cranes exhibit clear double-swing characteristics in actual working conditions, complicating anti-swing control. Most existing anti-swing control methods are based on a simplified single-pendulum model. The present paper puts forward a double-pendulum model for gantry cranes and proposes an adaptive hierarchical sliding mode control based on a neural network according to the actual working conditions. The use of a neural network and adaptive layered sliding mode control can effectively inhibit chattering, thus improving control performance and stability and achieving the goal of anti-shaking control, thus effectively inhibiting residual oscillation. This method has been demonstrated to be effective in achieving the objective of anti-shudder control, thereby effectively suppressing residual oscillation. Compared with hierarchical sliding mode control, the proposed method reduces the maximum residual oscillation angle of the hook and payload by approximately 80%. In comparison with the conventional sliding mode control, the maximum residual oscillation angle is reduced by approximately 84%. Furthermore, the control force amplitude is reduced to 5.23 N, representing decreases of 30.2% and 37.4%, respectively. These comparative results demonstrate the superior oscillation suppression. The system also shows a reliable performance against potential disturbances.