直接驱动X-Y平台递归神经网络控制仿真

Simulation of direct drive X-Y table based on recurrent neural network control

针对永磁直线同步电动机驱动的X-Y数控平台控制系统中存在的各种扰动,提出了一种基于自组织模糊递归神经网络的控制器设计方法,并利用切线-轮廓误差控制器对整体控制构架进行了整合式设计.该控制器融合了自组织模糊神经网络(SCFNN)和递归神经网络(RNN)的优点.仿真结果表明,所设计的控制系统对于参数的变化、外部的扰动等具有较强的抑制作用,减小了系统的轮廓误差,具有较强的鲁棒性.

A self-constructing fuzzy recurrent neural network （SCFNN） controller was presented for various disturbances existing in permanent magnet linear synchronous motor （ PMLSM ） driven X-Y table. The tangential-contouring error controller was adopted for the integrated design. The proposed controller combines the merits of self-constructing fuzzy neural network and recurrent neural network （RNN）. The simulation results show that the proposed controller can restrain the parameter variation and external disturbances. The method can reduce the contouring error and has strong robustness.