关键词: 桥式起重机, 防摇控制, 模型预测控制, 高斯回归建模, 数据驱动

Abstract: A bridge crane is very sensitive to the environment for its properties of nonlinearity,multi coupling and time variability of its parameters,especially for its underactuated property.A better control performance of bridge cranes can be guaranteed with anti-swing control algorithm based on a higher-accurate crane model,and the accuracy of crane model especially its nonlinearity can be improved based on data driven.A residual data model was identified by the Gauss modeling method,which learned the nonlinearity of cranes using the data of the actual input and output values and these values of the nominal model.The dynamic change law of crane system was continuously learned based on those streaming data to guarantee the accuracy of that residual data model.The anti-swing control algorithm was designed based on that data model of cranes to improve the robust of anti-swing control method of bridge cranes.The simulation and experimental results showed that using the proposed method could effectively improve the anti-swing control performance of bridge with better environmental adaptability and robustness.

Key words: bridge crane, anti-swing control, model predictive control, Gaussian regression modeling, data driven