Application of neural networks to a magnetic measurement system for mandibular movement

A magnetic measurement system for mandibular movement using neural networks is described in this paper. The position and orientation of a magnet which is attached to an incisor point can be estimated by using the magnetic field around the jaw. This problem is a type of inverse problem. It is difficult to design a fast and accurate measurement system using traditional methods for solving this inverse problem. The proposed system uses backpropagation neural networks to estimate the position and direction from the magnetic field. According to computer simulation, average errors are 7.2 μm for position and 0.002 degree for orientation. The processing time for one magnetic field sample is 4.2 ms. This results shows the proposed system is applicable to the measurement of the mandibular range of motion and conditions of occlusion