Nonlinear Calibration Model of a Triple Axis Tunnel Magneto Resistance Magnetometer

The accuracy of a triple axis tunneling magnetoresistance is limited by different scales, the zero offset of each axis, the nonlinearity of the sensitivity measurements, and nonorthogonality between axes. The typical tunnel magneto resistance (TMR) sensor’s sensitivity function is approximately by a hyperbolic tangent. We construct an objective function based on the principle of a constant total magnetic field. The calibration model parameters are solved via deep learning. Three typical TMR magnetometers are used for the nonlinear calibration of a triple axis TMR magnetometer. An experiment is designed that examines three different types of motion (cases), where the total magnetic field intensity around the typical TMR magnetometers remains constant. Experimental results show that, after adding a nonlinear correction term in the case of strong magnetic signal output, the root-mean-square (rms) error of calibration is reduced by a factor of nearly ten compared to the addition of a linear correction term. This result provides a straightforward approach for improving the nonlinear calibration performance of a triple axis tunneling magnetoresistance magnetometer.