Research On Integrated Attitude Determination System Based On MEMS/GPS For Micro UAV

Due to the advantage of low cost, low power consumption,high maneuverability and highconcealment, Micro Unmanned Aerial Vehicle(MUAV) has developed rapidly in the military andcivilian areas. Attitude and Heading Reference System(AHRS) can provide posture and sailinginformation, which is one of the essential airborne avionics system of aircraft to premise the flightsafety. The traditional AHRS is so expensive and heavy, that it’s not suitable for MUAV. The goal ofthe paper is to develop an AHRS for MUAV. The system is used to calculate the MUAV’s attitude inreal time, which provides measurements for attitude stabilization loop. Taking into account of theminiaturization, low power and low cost, the paper adopts the medium precision MEMS inertialdevice and GPS for the UAV’s attitude measurement. The main work is:For the question that MEMS inertial devices have larger error, more error source, the sensor errorsource of MEMS is analyzed, the model of the sensor error is obtained. For the characteristics ofMEMS sensors errors, we give the method of error calibration and compensation based on gravityfield excitation, constant speed excitation, and recursive least squares method of MEMSaccelerometers, MEMS gyros, and magnetometers respectively. This method is simple and effectivewith some certain values in engineering applications. It’s ideal for error calibration of low-costMEMS inertial devices.For the problem that the AHRS based on the gravity field and magnetic field vector observationsis susceptible to the motion acceleration, a method of using GPS derived accelerations is proposed toimprove attitude measurement accuracy under the high maneuverability. Combined with themovement characteristics of UAVs, we use GPS’s Doppler velocity measurement values asobservations and use JERK model to establish Kalman filter equations and then export the motionacceleration. By TRIAD algorithm, the vector observation equation is constructed using GPS derivedaccelerations, accelerometer and magnetometer. In order to solve the problems that the errorcovariance matrix will be singular and quaternion weighted averaging when quaternion is consideredas the state vector, UKF algorithm is proposed based on the Euler angles. We use the quaternion tocomplete the update process of Euler angles to solve the problem that Euler angles can not work forall attitudes.A kind of the quaternion complementary filter based on gradient descent algorithm is studied inthis paper. In order to achieve the correction of the quaternion error and the gyro drift, the Feedback correction amount in complementary filter under quaternion form is constructed by the gradientdescent method. For the question of the subjectivity and arbitrariness of filter parameters set in theapplication process, Particle swarm optimization algorithm is used to achieve filtering parametersoptimization. Under maneuvering, GPS acceleration is used to compensate accelerometer to obtainthe triaxial real component of Gravity field. Finally, the two algorithms studied in this paper areconducted in a semi-physical simulation experiment, the results show that the two algorithms bothhave good performance and are able to meet the requirements of MUAV.This paper uses the AHRS hardware platform based on the structure form of "ARM+MEMS+GPS（OEM board）"and completes the production of AHRS prototype. In ADS1.2integrateddevelopment environment, the Modular design ideas is used to complete AHRS software design.Finally, a physical verification and accuracy analysis are provided for this AHRS. The results showthat this AHRS proposed in this paper has the satisfactory attitude measurement accuracy both understatic and dynamic conditions.