Background: This paper presents a model-based bone milling state identification method that provides intraoperative bone quality information during robotic bone milling. The method helps surgeons identify bone layer transitions during bone milling.
Methods: On the basis of a series of bone milling experiments with commercial artificial bones, an artificial neural network force model is developed to estimate the milling force of different bone densities as a function of the milling feed rate and spindle speed. The model estimations are used to identify the bone density at the cutting zone by comparing the actual milling force with the estimated one.
Results: The verification experiments indicate the ability of the proposed method to distinguish between one cortical and two cancellous bone densities.
Conclusions: The significance of the proposed method is that it can be used to discriminate a set of different bone density layers for a range of the milling feed rate and spindle speed.
Keywords: artificial neural network modelling; bone density; force sensing; robotic bone milling; state identification.
© 2019 John Wiley & Sons, Ltd.