Navigation system is essential in estimating the orientation and translation motion of UAV. The system can utilize data from several sensors and devices, e.g. MIMU and GPS. The orientation of UAV can be precisely calculated by means of data from accelerometer and magnetometer when both sensors are in static state. Meanwhile, under dynamic conditions, the orientation can be more precisely predicted through the use of gyroscope sensor data. In order to attain a robust navigation system, a data fusion based on Kalman filter is built to estimate the orientation from the accelerometer, gyroscope, and magnetometer sensors. Moreover, for sake of achieving high accuracy, the filter will further correct the orientation by giving a higher weight to the data from accelerometer and magnetometer sensors when the UAV is in static condition. In the case of its position, the UAV position will be estimated by performing data fusion of MIMU and GPS data. Based on the experiments, it could be showed that data fusion based on Kalman filter provided more accurate results of orientation and position of UAV. The result of orientation based on gyroscope sensor data has an average error of 18.12°, while those obtained by accelerometer and magnetometer sensors data is 1.3°. Furthermore, by using the data fusion based on Kalman filter, the error of orientation predicted by combining data from accelerometer, magnetometer, and gyroscope will decrease to 0.87°.
Copyrights © 2021