Calibration of MEMS Accelerometers and Magnetometers for Rotor UAV Attitude Estimation

Abstract

Rotor unmanned aerial vehicles (UAV) usually adopt MEMS gyros, accelerometers and magnetometers to determine its navigation attitude. Because a MEMS gyro has a drawback of angle drift, its attitude data is often corrected by the data solved with accelerometers and magnetometers. This paper presents a static and scalar calibration method for accurate solving of attitude angles with MEMS triaxial accelerometers and triaxial magnetometers. Based on the facts that vector sum of triaxial outputs of accelerometers equals to the gravity acceleration, and vector sum of triaxial outputs of magnetometers equals to the geomagnetic vector, the error equations are established. Taking the sum of error squares as the objective function, a nonlinear least square method is applied to solve the optimal calibration parameters. A Kalman filter is used to suppress the random error of output signals of sensors. A precise triaxial turn table is used to vary the spatial attitude of the sensor module for data sampling. Output signals of sensors at 32 different attitudes are captured, and unknown calibration parameters are solved. It is found that the variation of the difference values between the attitude angles calculated with the calibrated parameters and the attitude angles indicated by triaxial turn table is around ±1o. It is proved that the proposed calibration method for MEMS triaxial accelerometers and triaxial magnetometers is accurate and feasible.