Neural Network-Based Nonlinear Stabilizing Control for 3-D Offshore Crane With Double-Pendulum Effect

Wave-induced ship motions pose great challenges to the design of control systems for offshore cranes, especially with double-pendulum effect and unknown system dynamics. In this paper, a neural network-based nonlinear stabilizing controller is proposed for 3D offshore cranes to accomplish boom positioning and payload swing-elimination under wave-induced ship roll and pitch motions. Critical and practical-oriented issues including double-pendulum effect, boom position limitations, actuator input dead zones, and unknown system dynamics are considered simultaneously. First, the dynamic model of 3D offshore cranes with double-pendulum effect is established by using the Lagrange’s modeling method. By combining the state variables with perturbation terms, new auxiliary variables are introduced for model transformation to simplify model analysis and controller design. Then, by analyzing the transformed model, the neural network is rationally designed to estimate the unknown system dynamics and input dead zones. In addition, barrier Lyapunov functions (BLFs) are employed to the controller ensuring that the boom operates within a safe range. The Lyapunov-based theory is utilized to rigorously prove the stability and convergence of the designed control system. Hardware experiments are elaborately designed to verify the performance of the proposed method in terms of effectiveness, robustness, and anti-disturbance. Note to Practitioners—This paper studies the stabilizing control problem of offshore crane systems. In practice, the double-pendulum effect between the payload and hook is evident, and it is necessary to simultaneously lift and rotate the boom to position and stabilize the payload smoothly and accurately under wave-induced ship motions. However, existing studies typically oversimplify the model of offshore cranes, neglecting crucial factors such as their 3D characteristics, the double-pendulum effect, and the intricate wave-induced ship motions. This limits the application of the corresponding control methods in practice. Moreover, most of the existing control methods ignore unknown model dynamics, input dead zones, etc., which is not favorable for practical applications. To address the above problems, this paper proposes a neural network-based nonlinear stabilizing controller, which can accomplish the stabilization control for 3D offshore cranes with double pendulum effect in the presence of unknown dynamics and input dead zones. The effectiveness of the proposed method is validated on the self-built hardware platform.