二维直线电机平台的双神经网络前馈解耦控制

Feedforward-Based Double Neural Network Decoupling Control for Two-Dimensional Linear Motor

针对二维直线电机系统中的直线电机本身存在的非线性以及XY两轴之间互相影响产生的耦合对控制精度的影响,提出了一种基于前馈结构的双神经网络解耦控制方法。根据二维直线电机两轴间的耦合结构和前馈控制结构,对应搭建了双神经网络模型,设计一种双神经网络的解耦控制策略,以消除二维直线电机两轴间的非线性强耦合对系统控制精度的影响。仿真结果表明:提出的双神经网络前馈解耦控制方法的有效性,并在起始阶段的跟踪方面表现出良好的性能。

For the two - dimensional linear motor, the nonlinear of linear motor and the coupling between X - axis and Y - axis influence on the precision of two - dimensional control system. To solve the problem, a feedforward - based double neural network decoupling control is proposed. The double neural network model is obtained by combi- ning the coupling structure between two axes of the two - dimensional linear motor and the feedforward control decou- pling method. A kind of the dual neural network decoupling control strategy is designed to eliminate the influence of the strong nonlinear coupling between two axes of two - dimensional linear motor on the control precision. Simulation results show that the proposed feedforward - based double neural network decoupling control is effective, and has a good performance in both decoupling and tracking in initial stage.