Research On Mins/Mag/Vision Integrated Navigation System

With the continuous improvement of MEMS devices,their performance and reliability continue to improve while the cost,volume and power consumption continue to reduce,MEMS inertial measurement unit(IMU)can gradually meet the needs of various applications in the field of autonomous navigation,and has been widely used.However,when MEMS inertial measurement unit is used alone to provide navigation information,the problem of divergence of errors cannot be avoided in long-term applications.It is still necessary to provide more accurate navigation information combined with other devices.Among them,geomagnetic navigation has the characteristics of long-term stable global coverage.The cost of vision sensor is low,and it can complete the task of location velocity measurement by detecting the feature changes of continuous images.In this paper,a kind of integrated navigation scheme of MIMU-MAG-VISUAL is designed.The main research contents are as follows:The overall scheme of the system is analyzed and designed,and the error model and calibration scheme of accelerometer,gyroscope and geomagnetic sensors are studied to improve the accuracy of sensor data in order to improve the navigation accuracy.The calculation method of attitude in inertial navigation system is given.The selection mode of navigation coordinate system and carrier coordinate system is set,and three attitude angle description methods with attitude updating methods under various representation methods are given.The mathematical transformation between the three representation methods is given.By analyzing the advantages and disadvantages of several representation methods,the application mode of attitude calculation in this paper is determined.Two multi-sensor fusion filtering methods are designed to overcome the shortcomings of micromechanical devices with low accuracy and large gyroscope drift compared with traditional high-precision devices.One is attitude reconstruction based on gradient drop method,the other is attitude filtering based on complementary filter.Due to the problem of inaccurate attitude measurement of accelerometer and magnetometer under high dynamic condition,an adaptive algorithm of filtering parameters is designed to improve the accuracy of attitude calculation.Gait detection is designed,and zero speed correction is introduced to correct the state error of the system under zero velocity motion.The zero velocityinformation and the attitude output obtained by visual navigation are used as the observation,and the attitude,velocity and position errors are corrected by kalman filter.Finally,the swing experiment is designed to test the performance of the navigation algorithm.