Two-Phase Attractors Reaching Law-Based Adaptive Control With Neural Network for Overhead Cranes

The article proposes a two-phase attractors reaching law method with neural network for underactuated overhead cranes to tackle control issues arising from system uncertainties. In the reaching law design presented, the power value is adjusted based on changes in the sliding mode variable, facilitating rapid convergence of the sliding mode variable both when far from and approaching equilibrium. A precise expression for the convergence time of the sliding mode variable is derived by computing two distinct convergence times corresponding to different reaching stages. Additionally, the study develops an adaptive neural network capable of approximating nonlinear uncertainties inherent in the system, such as unmodeled dynamics and uncertain parameters. This adaptive neural network reduces the reliance on precise system modeling. The effectiveness of the proposed control scheme is rigorously validated through mathematical analysis, demonstrating the stability of the system. Moreover, experimental validation through comparative tests on overhead cranes shows the effectiveness of the proposed control strategy.