六自由度虚轴机床神经网络位置伺服控制

针对永磁直线伺服电机（PMLSM）驱动的六自由度虚轴机床，提出利用基于神经网络的位置伺服控制器对电机动子加速度进行控制的方案。为抑制对电机动子加速度影响较大的杆问耦合扰动的影响，将负载扰动和杆间耦合扰动整合为一个对象，并等效为电机动力学参数的变化，利用观测器进行在线观测，结合神经网络位置伺服控制器进行调整，从而较好地提前抑制了速度与位置误差。所设计的神经网络位置伺服控制器结构简单，权值调整方便，计算量小、响应速度快。仿真实验表明该方法对虚轴机床各杆的位置、速度、加速度实现了精确的跟随控制，系统具有很强的鲁棒性。

Position Servo Control Based on Neural Network for 6-DOF Virtual-axis Machine Tool

Aiming at 6-DOF virtual-axis machine tool driven by permanent magnetic linear synchronous motor (PMLSM),a control scheme on rotor acceleration was presented based on neural network position servo controller. In order to overcome the influence of the coupling force between links, which have a strong affection to rotor acceleration, the load disturbance and coupling force have been integrated and was regarded as the dynamic parameters' variation of PMLSM, and the equivalent parameters were online observed by observer and adjusted by neural network position servo controller, the errors of the speed and position were decreased ahead. The controller has the advantages of simple structure, convenient weights adjusting, small calculation and fast response, etc. The simulation results verified that the tracking control for the position, speed and acceleration of each link was attained, and the system is robust to uncertain factors.