Bone Milling: On Monitoring Cutting State and Force Using Sound Signals

Abstract Efficient monitoring of bone milling conditions in orthopedic and neurosurgical surgery can prevent tissue, bone, and tool damage, and reduce surgery time. Current researches are mainly focused on recognizing the cutting state using force signal. However, the force signal during the milling process is difficult and expensive to acquire. In this study, a neural network-based method is proposed to recognize the cutting state and force during the bone milling process using sound signals. Numerical modeling of the cutting force is performed to capture the relationship between the cutting force and the depth of cut in the bone milling process. The force model is used to calibrate the training data to improve the recognition accuracy. Wavelet package transform is used for signal processing to understand bone-cutting phenomena using sound signals. The proposed system succeeds to monitor the bone milling process to reduce the surgical risk. Experiments on standard bone specimens and vertebrae also indicate that the proposed approach has considerable potential for use in computer-assisted and robot-assisted bone-cutting systems used in various types of surgery.