matlab中mag,Magnetometer calibration coefficients

Because a magnetometer usually rotates through a full range of 3-D rotation, the ideal

measurements from a magnetometer should form a perfect sphere centered at the origin if the

magnetic field is unperturbed. However, due to distorting magnetic fields from the sensor

circuit board and the surrounding environment, the spherical magnetic measurements can be

perturbed. In general, two effects exist.

The soft-iron effect is described as the distortion of the ellipsoid from a sphere

and the tilt of the ellipsoid, as shown in the left figure. This effect is caused by

disturbances that influence the magnetic field but may not generate their own magnetic

field. For example, metals such as nickel and iron can cause this kind of distortion.

The hard-iron effect is described as the offset of the ellipsoid center from the

origin. This effect is produced by materials that exhibit a constant, additive field

to the earth's magnetic field. This constant additive offset is in addition to the

soft-iron effect as shown in the figure on the right.

The underlying algorithm in magcal determines the best-fit

ellipsoid to the raw sensor readings and attempts to "invert" the ellipsoid to produce a

sphere. The goal is to generate a correction matrix A to account for

the soft-iron effect and a vector b to account for the hard-iron

effect. The three output options, 'eye', 'diag' and 'sym' correspond to three

parameter-solving algorithms, and the 'auto' option chooses among these three options to

give the best fit.