Magnetic motion measurement using BPNN

Abstract

It is important in various fields of industrial instrumentation and bioinstrumentation to measure position and direction in dimension of object. It is preferable that bioinstrumentation is non-contact. But technique for making measurement handy and highly accurate is a little when measurement thing exists in the part of in vivo and is hidden. Because body tissue is percolation to magnetic field, it seems that method of using magnetic field is the most suitable for such a motion measurement. In this research, magnetic motion measurement system that uses magnetic field is examined. Mandibular motion measurement that uses magnetic field is examined as an application example of applying to these conditions. Because jawbone has been covered with skin, measurement is difficult. But mandibular motion can be measured because magnetic field is percolation to body tissue. Here, magnetic flux density can be measured from position and direction of magnet. But, problem to calculate position and direction of magnet from magnetic flux density is a kind of a reverseproblem. Generally, this problem cannot be solved. This reverse problem can be solved by BPNN. Therefore, position and direction of magnet can be presumed from magnetic flux density of magnet by using BPNN. BPNN is abbreviation of Back Propagation Neural Network. BPNN can approximate a complex nonlinear functions. Input of BPNN is magnetic flux density distribution, and output is a position and direction of magnet. Input-output relation of neural network can be trained by using training patterns of a lot of limited numbers.