Improving Azimuth Estimation Accuracy For INS Based On MEMS Sensors

The accuracy of azimuth plays a decisive role for the accuracy of strapdown inertial navigation system(SINS).Due to the limited precision of the MEMS gyroscope,it is difficult to estimate the accurate azimuth for a long time.It is also necessary to study how the SINS based on MEMS sensors can achieve the high accurate of azimuth for a long time.In view of the above problems,this thesis uses some auxiliary information to calibrate the deviation of azimuth,which is mainly from the following three aspects:(1)The whole design for improving the accuracy of azimuth estimationThe factors that affect the azimuth estimation accuracy are analyzed and studied.Aiming at the azimuth error that has been generated,we design the whole calibration scheme.The real-time magnetic interference assessment is carried out on the collected data environment,and the different schemes are selected according to the different degrees of magnetic interference.When the geomagnetic field is not significantly affected by interference or is only occasionally interfered,the azimuth error is calibrated based on the combination of complementary filter and Kalman filter.When the geomagnetic field is affected by interference seriously,azimuth error calibration with the scene constraints will be adopted according to the result of human motion mode.(2)The algorithm of azimuth calibration based on Kalman filter and complementary filterWhen the magnetic field is not disturbed obviously,the algorithm that combines the Kalman filter with the complementary filter is proposed.Firstly,the optimal estimation of quaternion is obtained by the fusion of the electronic compass and the gyroscope through the Kalman filter.Then the complementary filter algorithm is used to compensate the drift of the gyroscope and get the quaternion that corrected.The quaternion obtained from these two times will go through the Kalman filter algorithm again to make the second optimal estimate for the quaternion.And then we get the accurate azimuth.The azimuth accuracy of no filtering algorithm,Kalman filter algorithm,complementary filter algorithm and the fusion algorithm are compared in the experiment.The experimental results show that the algorithm can effectively solve the error within 1°/h.(3)The calibration of azimuth error with the scene constraints according to the result of human motion mode.When the geomagnetic field is seriously interfered for a long time,the electronic compass will not be able to accurately determine the azimuth information,then the azimuth error calibration with the scene constraints will be adopted according to the result of human motion mode.When pedestrians remain still in the elevator,pedestrians must be in a vertical motion state at this time,and then we can use the vertical component to keep constant observation and correction.When pedestrians are walking in a straight line,do subtraction between the adjacent zero speed intervals of the vertical position and the heading angles,and we can get the figure which can be used as the correction of Kalman filter.When pedestrians are walking on stairs,the height of each step is usually certain.The length between the two adjacent steps can be used as the correction of Kalman filter.The experimental results show that the method is feasible and efficient,and can achieve the error within 1°/h.