在分析专家系统和人工神经网络特点的基础上,提出了一种基于人工神经网络的零件定位基面选择方法,分组建立了BP网络模型,并根据零件特征提出了网络的输入和输出编码方法;针对BP网络的特性,在训练中采用了变步长和变精度的方法来加快收敛速度,并给出了实例 Based on the analysis of the characteristics of expert system and artificial neural network, this paper proposes a method of selecting base of positioning of parts based on artificial neural network. The BP network model is established by grouping and the input and output coding methods of network are proposed according to the features of parts. Aiming at the characteristics of BP network, the method of variable step size and variable precision is adopted in training to speed up the convergence speed and gives an example