基于神经网络的自由曲线插补研究

Study on Curve Interpolation Arithmetic Based on Neural Network

介绍了利用神经网络的非线性逼近能力和自学习能力,对未知表达式的曲线进行辨识,并利用从曲线的神经网络辨识模型中得到的数据,建立一个数控系统的神经网络插补控制器,实现对未知表达式曲线的插补。通过理论研究和仿真试验表明,这种方法能够较好地完成这类曲线的插补。

This paper introduces the strong abilities of nonlinear processing and self - learning of neural networks. By using these abilities, an expression of the unknown - formula curve is simulated. Then the neural networks model for CNC interpolation is built by using the data points that get from the simulated expression. The research and the simulation calculation demonstrate that this method can solve the interpolation of these unknown-formula curves.