Hilbert-Huang Transform based state recognition of bone milling with force sensing

Bone milling is one of the most common operations in various kinds of orthopedical surgeries, such as laminectomy surgery. For safety issue and efficacy, it is very important to recognize the states in milling operation. In this paper, an approach to recognize the states of bone milling is proposed, which identify the cortical tissue layer and cancellous tissue layer. Hilbert-Huang Transform (HHT) based on Empirical Mode Decomposition (EMD) is used to analysis and extract the features of the interactive force in milling operation. The instantaneous amplitude of the Intrinsic Mode Functions (IMF) are combined by means of linear weighting method to obtain one comprehensive evaluation index. The feature vector of the index consists of average amplitude, kurtosis, crest factor and average remaining of EMD. With the feature vector, states of cortical and cancellous layer in milling process are recognized based on Support Vector Machine (SVM). Finally, the milling experiment with pig scapula is performed to show the effectiveness of the proposed approach.