Algorithm Design And Experimental Research On Attitude Of Micro-aircraft Based On MEMS Sensor

Compared with only the GPS mounted on a micro-aircraft,we can obtain the accurate micro-aircraft attitude with micro-MEMS sensors also equipped at the same time.MEMS gyroscope sensor,accelerometer,magnetometer were separately measured angular velocity,acceleration and magnetic force values.Compared with the traditional navigation sensors,MEMS sensors have more and more applications due to their advantages of small size,low cost and convenient carrying.Due to cumulative errors,such as the deterministic drift and the drift of randomness of MEMS motion sensors,the attitude error is getting larger and larger,causing MEMS sensors not directly suitable for mounted on a micro air vehicle.Therefore,how to compensate for the error of the sensor to improve the measurement accuracy and suit for being mounted has become the focus of the research on MEMS sensors.At present,there are two ways to improve MEMS sensor accuracy,one is to reduce the accumulated error by improving the level of technology,but obviously this method is not suitable due to the characteristics of gyro cumulative error,and the high cost of the research process.Another is to create the "black box" model of attitude calculation.It can improve the accuracy of attitude due to its low cost and short cycle.The "black box" model of attitude calculation of the micro air vehicle involves in the algorithms such as inertial navigation,error modeling and Kalman filtering.The main work of this thesis is as follows:(1)This paper introduces the inertial navigation at first,mainly introduces coordinate system and the transformation matrices between the geographic coordinate system and the carrier coordinate system,and introduces several theories of attitude calculation,especially the quaternion algorithm,which is also the core algorithm of attitude calculation of micro air vehicle.In addition,we introduce Kalman filter,a very important algorithm in the model,and explain the Kalman filter principle,the state process,the measurement process and the covariance.(2)The JY901 module of the motion sensor is highlighted,and the principle and characteristics of the gyroscope,accelerometer and magnetometer of the module are described.The initial attitude of the "black box" model of attitude calculation is deduced based on the accelerometer and the magnetometer,and the maximum theoretical error is analyzed.The initial attitude of the micro aircraft is an important process whose accuracy affects the attitude of micro-aircraft during motion process to a large extent.This provides initial attitude angle for the follow-up the "black box" model of attitude calculation.(3)The motion attitude calculation of micro-aircraft is the core of the "black box" model.The motion attitude of the aircraft uses the data of the gyroscope.The algorithm of the moving attitude includes time series analysis,inertial navigation algorithm and Kalman filter algorithm,finally obtaining the accurate angular velocity output.Motion attitude cover both static and dynamic parts,which models static and dynamic process of MEMS sensor respectively,and determine process noise covariances,measurement noise covariances,error-modeled state transition matrices through theoretical calculations and extensive experimental data.These parameters act as the interface to input of the "black box" model of attitude calculation,and we can obtain more accurate attitude using the model.And we can use the turntable experiment to verify its accuracy,which proves higher accurate attitude of the output by using the "black box" model of attitude calculation.And then implements micro-aircraft attitude algorithm on the ARM.(4)The error sources of the process noise of the gyroscope are analyzed by Allan variance.And the errors before and after Kalman filtering are compared to illustrate the correctness of the filtering.This work laid the theoretical and experimental foundation for the high-precision attitude control of micro-aircraft.