Research On An Attitude Controller For Four-rotor UAV Based On MARG Sensors

In recent years,with the rapid development of MEMS technology and embedded processor,the MARG sensor module which integrate accelerometer,magnetometer and gyroscope with small size,light weight,low power consumption have appeared.The sensor module can not only be used for airborne four-rotor attitude detection,but also can be used to design a new ground attitude remote control.Based on MARG sensor module and multi-sensor data fusion technology,this paper designs a novel attitude controller,which realize the attitude control of the four-rotor.In order to ensure the accuracy of output data of MARG sensor module,the error sources of accelerometer,magnetometer and gyroscope in MARG sensor module are analyzed,and the error models of accelerometer,magnetometer and gyroscope are established.The MARG sensor module is fixed on the three-axis manual turntable,the angle of the three-axis turntable is adjusted and the output data of the sensor module is acquired under the whole attitude.To complete the static calibration of three-axis accelerometer and three-axis magnetometer,the non-linear least squares method is used to obtain the scale factor matrix and zero bias coefficient,And then complete the rate calibration of the three-axis gyroscope through speed turntable,the three-axis gyroscope scale factor matrix and the bias coefficient are obtained.The output datas of MARG sensor module are corrected by the respective scale factor matrix and zero bias coefficient calibration.In order to ensure the accuracy of the attitude estimation,the complementary filtering and extended Kalman filter are used as the fusion of the attitude angle datas.Complementary filter fusion algorithm can fuse the resolution data from gyroscope,accelerometer and magnetometer,which use accelerometer and magnetometer low frequency effective output to suppress the drift of the gyroscope according to the frequency response characteristics of different sensor modules.and the bias of the attitude estimation is reduced.The extended Kalman filter fusion algorithm corrects the output of the gyroscope using the difference between the estimated value of the accelerometer and the magnetometer and the calibrated measurement values.The validity of the algorithm is verified by comparing the solution values and the fusion values of the input datas of different sensor types in the Matlab environment.The fusion algorithm of multi-sensor data improves the attitude measurement accuracy and resistance to vibration and shock.Based on the MPU6050-HMC5883 L sensor module and the STM32F407 microcontroller,a four-rotor attitude controller was designed and manufactured.The fusion algorithm is implemented by C language programmed in STM32F4.The postfusion attitude angle value is sent to the four-rotor signal receiver via the 2.4G wireless module as the control value.The four-rotor is controlled in real time.The host computer also records the attitude angle transmission datas and the attitude datas of four-rotor.The experimental results show that the attitude curve of the four-rotor can follow the attitude curve of the MARG sensor module in real time.The attitude controller can flexibly control the attitude of the four-rotor,and verify the feasibility of the attitude controller and the data fusion algorithm of attitude angle.