Moreover, the feature physical parameters are selected in view of four typical conditions of shape-milling tool, namely acoustic emission signal, cutting force signal and power signal.

• 其次，针对成形铣刀常见的典型的四类状态，选取了与刀具状态密切相关的特征物理量作为神经网络的学习样本，即声发射信号、切削力信号、功率信号等。