Constrained Total Least Squares Algorithm Applied to Three-axis Magnetometer Calibration

Strapdown three-axis magnetometer is becoming increasingly used in navigation systems and in biomedical applications. A calibration must be performed before the magnetometer can be used, because its outputs are often corrupted by various magnetic field sources. This paper presents a new calibration algorithm based on constrained total least squares (CTLS) to determine different calibration parameters such as sensor non-orthogonality, scale factors, offsets, hard iron and soft iron errors. The basic idea is to convert the calibration problem into an unconstrained nonlinear optimization problem. An iterative technique based on Newton's method is applied to give a numerical solution. Simulation results show that the proposed CTLS algorithm gives more accurate solutions than compared algorithms.