交流永磁直线伺服系统的神经网络--滑模双自由度控制

文章针对直接驱动的交流永磁直线伺服系统，提出一种将非线性神经网络控制和滑模控制相结合构成的双自由度控制策略。该控制策略解决了直线伺服系统跟踪性能的鲁棒性能之间的矛盾。采用滑模控制方法设计输入控制器，保证系统对给定的快速跟踪性能；输出反馈控制器采用神经网络来实现，对系统参数变化和阻力扰动（包括直线电机端部效应引起的推力波动）进行很大程度的抑制。并可以消除扰动引起的滑模控制抖振对系统稳态性能的影响。同时，滑模控制的快速性又能大大加快神经网络的收敛速度。仿真实验结果表明该方案在保证伺服系统的快速性同时，对系统参数变化和阻力扰动具有很强的鲁棒性，大大提高了直接驱动系统的伺服精度。

Two-degree-freedom Control with Neural Network-sliding Mode for AC Permanent Magnet Linear Servo System

A two degree of freedom control scheme based on combining nonlinear neural network(NN) control and sliding mode control for a direct drive AC permanent magnet linear servo system is presented in this paper.The control scheme solves the conflict between tracking and robustness of the linear servo system.The input controller is designed based on sliding mode control method,which can ensure the system have fast tracking characteristic to the command.The output feedback controller is realized by a neural network,which can suppress parameter variations and resistance perturbation(including the thrust undulation produced by end effect of linear motor )greatly,and can effectively eliminate the chattering of sliding mode control which could influence the steady state performance of the system. Simultaneously,the convergence speed of NN can be increased greatly by the speediness of sliding mode control.Some simulation results show that this control scheme not only can enhance the track command ability and the robustness of the linear servo system,but also has strong robustness to parameter variations and resistance perturbation.The servo precision of direct drive system is improved greatly.