New Compensation Method of Magnetometer Time-varying Bias for UAV
Heekwon No, Seoul National University and SNU-IAMD, Republic of Korea; Am Cho, Korea Aerospace Research Institute, Republic of Korea; Changdon Kee, Seoul National University, Republic of Korea
Location: Grand Ballroom F
Date/Time: Thursday, Feb. 1, 11:00 a.m.
A new method to compensate the time-varying bias of the magnetometer within the electric-powered UAV is proposed in this study. The Earth's magnetic field is a useful physical quantity, which provides attitude information to the navigation system. Since the electric current generates a magnetic field, distortion of the measured magnetic field occurs when there is a current flowing in the vicinity of the magnetometer. In the case of an electric-powered UAV, a large current is used for driving an electric motor used as a propulsive device. Therefore, the influence of the current on the magnetometer is serious and must be eliminated before use. The magnetic field generated by the current is proportional to the amount of the current flow. Therefore, using the measurement of the current sensor, it is possible to compensate the influence. However, the motor of electricpowered UAV is driven by a signal of pulse width modulation (PWM) type, so there is a lot of noise in the current measurement. Accordingly, if the measurement of magnetometer is compensated using the noisy measurement of the current sensor, then the bias error can be compensated, but the noise is greatly increased. In this study, the current measurement and the throttle input are combined to estimate the current of low-noise, which compensates the effect of the current without the increase of noise. Using the relationship between the current and the throttle input, the current of low-noise is estimated. This current estimate is used to compensate the time-varying bias error of the magnetometer, and reduces the noise exists in the compensated measurement of the magnetometer. Test results show that not only the bias error but also the noise is compensated effectively, in contrast to the previous method.