基于AR特征的刀具状态识别方法

Research on the tool condition reorganization based on AR feature extraction

通过分析切削过程刀具产生的振动信号的特点,引入自回归(AR)模型来表征刀具切削过程的工作状态;并利用隐Markov模型(HMM)对经AR模型处理后得到的特征向量(AR系数)和由FFT得到的特征向量(幅值谱)进行比较.结果表明:对于切削过程产生的振动信号采用AR模型得到的特征参数比由FFT得到的幅值谱更能有效地表征这一过程中刀具的工作状态.

By analyzing vibration signals generated by tools, the work condition in cutting process can be described by autoregressive （AR） model. By using hidden Markov model （HMM）, a comparison between AR coefficient feature vector generated by AR model and amplitude spectrum from FFT was made. The result showed that it was more effective to describe the work condition of tool in cutting process by using AR model than by FFT method.