Name: SÉRGIO SILVA MUCCIACCIA
Publication date: 22/08/2017
Advisor:
Name |
Role |
|---|---|
| ANSELMO FRIZERA NETO | Advisor * |
| EVANDRO OTTONI TEATINI SALLES | Co-advisor * |
Examining board:
| Name |
Role |
|---|---|
| ANSELMO FRIZERA NETO | Advisor * |
| EVANDRO OTTONI TEATINI SALLES | Co advisor * |
| PATRICK MARQUES CIARELLI | Internal Examiner * |
Summary: The measurements obtained from magnetometers are sensitive to disturbances and errors, requiring a calibration method that can considerably improve accuracy. The ellipsoid fitting is one of the most widely used methods for magnetometer calibration, but most
algorithms use iterative methods, causing runtime and convergence problems. As an alternative, a direct algorithm based on the method of least squares using the algebraic distance metric is proposed.
This present work presents an algorithm of calibration of magnetometers and its use in a system of calibration and fusion of data of magnetometers, accelerometers and gyroscopes based on a Kalman filter forming an inertial sensor able to obtain its orientation in the space. Computational simulations and tests with real data show that the calibration algorithm eliminates almost all the linear errors while performing much faster than traditional algorithms. Measurements of a magnetometer calibrated with the proposed algorithm are used in conjunction with measurements from accelerometers and gyroscopes to form an inertial measurement unit (IMU) using a simple Kalman filter. The complete system worked as
expected and the test results indicate that the magnetometer calibration algorithm is suitable for use in an IMU being more than ten times faster than traditional algorithms and presenting similar accuracy.
