基于神经网络的测量数据实时轨迹补偿方法

Real-time Track Compensation Method for Measurement Data Based on Neural Network

目前针对轮廓、倒边的制作过程,因为受到诸如夹持情况下的不均匀受力等的外力影响,极易发生轮廓扭曲。加工的标准是在规定的范围内确定每次的切割量,不然就会导致不均匀的倒角宽度。文章分析的实时轨迹补偿方式是实现测量数据基础上的,首先测量计算测量点位置形变,通过获取的测量点位置神经训练模型和形变量。是一种基于测量数据实现实时轨迹补偿的方法,将轨迹点在理论轨迹上的位置引入神经网络模型,轨迹点位置的变形量等同于输出函数值,通过叠加变形量和理论轨迹获取具体位置量,该方法可以有效地避免轨迹的拐点,易于得到补偿值,更加准确反映工件变形量。

The current production process of contour and inversion edge is easily distorted due to the external force such as the uneven force under the clamp condition.The standard of processing is to determine the amount of each cut within the specified scope,otherwise it will lead to an uneven chamfer width.The real-time trajectory compensation method is realized based on measurement data,the technique in the paper is the method of realizing real-time trajectory compensation,introducing the deformation of the track point to the output function value of the track,easily get the compensation value and more accurately reflect the deformation of the workpiece.