Research on Real-Time Anti-Sway and Positioning of Quay Cranes Trolley Based on Deep Reinforcement Learning

To address the nonlinear under-actuated sway and precise positioning challenges caused by the flexible wire rope of quay-crane trolleys, conventional electronic anti-sway methods suffer from insufficient robustness due to their reliance on accurate mathematical models, a real-time anti-sway and precise-positioning control strategy based on deep reinforcement learning (DRL) is proposed. A co-simulation platform is developed through secondary development of Adams, and the state space modeling is optimized using the Long Short-Term Memory–Proximal Policy Optimization (LSTM-PPO) algorithm to achieve closed-loop control in dynamic environments. The experimental results show that the positioning error of the quay bridge trolley is ≤ 5cm, and the residual swing angle within 3s is