永磁同步电动机自适应神经网络IP位置控制器

针对永磁同步电动机伺服系统（PMSM）存在结构复杂以及系统性能受不确定性影响严重等问题，在IP位置控制系统上，提出了神经网络补偿器的方法。该方法利用参考模型和系统输出的偏差在线训练神经网络的权值，有效克服了参数变化和负载扰动等不确定因素的影响，而且控制方法简单，动态快速性好。仿真结果表明，采用神经网络补偿器的方法能明显改变系统的品质，增强系统的鲁棒性。

IP position controller based on adaptive neural network for PMSM

The permanent magnet synchronous motor(PMSM) drive system has many drawbacks,such as the complicated structure and the performance significantly influenced by various uncertainties,etc.Based on an IP position controller,a neural network compensator is proposed.The neural network weights are online trained by the error of the reference model and the actual system output in order to overcome the uncertainties of parameter variations and load disturbances.The method has simple control algorithm and fast dynamic performance.The simulation experiment results verify the system can improve performance of system effectively and enhance the robustness.