Abstract
Drilling is one of the most common forms of tissue removal procedures, and drilling to a desired depth contributes to avoid injury to the soft tissue beyond and ensure implant stability. The deformation of the human musculoskeletal system has been a common problem in many drilling processes, making it difficult to achieve accurate estimation of the drilling depth. To remedy this problem, a dynamic model is presented to describe the relationship between the axial vibration of the drill and the feed rate. During drilling process, the amplitude of the main harmonic is estimated from the high-frequency component of the acceleration signal, while the short-time integral of the low-frequency part is calculated. Both the initial contact of the drilling tool to the bone and breakthrough are identified by comparing either the harmonic amplitude or the short-time integral. The harmonic amplitude is mapped to the data from a non-contact position sensor tracking the feed rate of the drill. Multiple drilling experiments on both a handheld device and a robotic cutting system demonstrated the effectiveness, stability and accuracy of the method when estimating depth. The mean maximum error for drilling depth estimation is less than 15% of the simulated bone thickness when using the handheld device, while the mean maximum error is less than 5% for the robotic cutting system.