Research On Algorithm Of Attitude Heading Reference System Based On MEMS Sensing Technology

Tech level in the 21st century to promote the progress of microelectronics technology and micro-machining level, MEMS inertial sensors have emerged at this time and gradually become the world focus in the field of inertia research. The AHRS constructed mainly by MEMS inertial measurement unit and it has advantages of small size, light weight and low cost that made it widely used in many fields, such as robots, unmanned aerial vehicles,automobiles, ships, submarines, communications equipment and bio-medicine.Micro inertial system can not complete the job of attitude mesuarement independently because of the large systematic error due to its low accuracy, the error accumulation over time and other factors. The AHRS consisted of the MEMS inertial measurement unit,magnetometer and GPS, and it fused three types of information.This paper describes the common reference coordinate system used by AHRS and the magnetometer measuring principle. And then we study the analytic initial alignment method that made use of the output information of MEMS accelerometers and magnetometers to work.In addition, we introduced and contrasted the various attitude updating algorithms of MEMS AHRS and choose the quaternion algorithm to complete attitude updating.We analyzed the MEMS gyroscopes, MEMS accelerometers and magnetometers errors and established the output error model for MEMS gyroscope, MEMS accelerometer and magnetometer based on the basic theoretical study. Then, we designed the gyroscope angular increment calibration program, the accelerometer six position calibration program and the magnetometer error calibration program based on the assumption of ellipsoidal according to the error models. Then we calculate the error parameters of each component and compensated them to lay a solid foundation for future research.During the study of AHRS algorithm we first established the nonlinear error equations of MEMS inertial system based on large misalignment angles. Under GPS signal, we used EKF filtering algorithm for nonlinear systems to fuse the information of MIMU, GPS and magnetometer. On this basis, we used the errors between GPS location, GPS speed, magnetic heading and inertial system own position,speed,heading as the measurement of the EKF to estimate the system attitude. The simulation results verified the effectiveness of the algorithm.In the absence of GPS signals, we were use of the improved adaptive Kalman filter algorithm based on quaternion error to combine MIMU information and Magnetometer information for attitude estimation. At this algorithm we used the gyroscope for attitude solver while accelerometer and magnetometer for the assistance. The second-order measurement update algorithm for accelerometer and magnetometer and the adaptive algorithm compensating external acceleration were proposed. At last, we proved the rationality of the algorithm by the Matlab simulation result.On the basis of simulation,laboratory static test,outdoor automotive test and Songhua River test were designed. Among them, the static test results are more satisfactory; the solved attitude for outdoor automotive test had a certain accuracy because of the car had been driving on a smooth road in the car testing process; Songhua River test results are further explained the combined method for MIMU, GPS and magnetometer can tracking pose effectively at low dynamic conditions.