基于迭代学习控制的划片机高精度运动误差补偿

集成电路封装技术快速发展,作为电子元器件封装关键设备的划片机精度也日益提高,为顺应这一趋势,针对划片机重复加工的执行过程特点,设计一种基于数据驱动的神经网络迭代控制方法。该方法通过不断修正控制输入,使得跟踪误差逐渐减小,提高重复定位精度。通过仿真分析得出,神经网络迭代控制方法基于划片重复运行的特点,仅利用输入输出数据,完成对零件加工过程的改善,可有效提高运动精度,实现了误差补偿。

Error Compensation for High-Precision Motion of Dicing Saw Based on Iterative Learning Control

With the rapid development of IC packaging technology,the precision of dicing saw,which is the key equipment for packaging electronic components,is also improving day by day.In order to comply with this trend,a neural network iterative control method based on data-driven is designed according to the characteristics of repetitive processing of dicing saw.In the method,the tracking error is gradually reduced and the repeated positioning accuracy is improved by constantly correcting the control input.Through simulation analysis,it is concluded that the neural network iterative control method is based on the characteristics of repeated dicing operation,and only uses the input and output data to improve the machining process of parts,which can effectively improve the motion accuracy and realize the error compensation.