Neural network sliding mode anti-swing control for tower crane

Aiming to the swing of load in the process of transportation, the dynamic model of tower crane system is analyzed. Based on genetic algorithm, a new method of neural network sliding mode anti-swing control is proposed for tower crane. The outputs of neural networks are used to approach the uncertainties of the positioning subsystem and the anti-swing subsystem. The parameters of the controller are optimized with genetic algorithm (GA), which can improve the convergence speed of parameters. This method weakens the system chattering which sliding mode control (SMC) bring out, enables the system to have good dynamic performances and enhance the quality of control system. The simulation results show that the feasibility and effectiveness of the method.