Research On Fusion Method Of MARG Sensor Based On AR Kalman Filter

The heading system based on MEMS gyroscope,MEMS accelerometer and three-axis magnetometer has the advantages of low power consumption and low cost,and has broad application prospects in the fields of marine unmanned boats and small aircraft.The low-cost attitude and attitude system solves the attitude of the carrier through the MEMS gyroscope and the MEMS accelerometer,and uses the magnetometer information to compensate the zero drift of the gyroscope.When the carrier moves in an environment with a lot of magnetic interference,the output of the magnetometer will have a large error,resulting in an error in the attitude resolution of the AHRS.In order to solve the problem of carrier attitude calculation in the magnetic interference environment,the main research work of the subject is as follows:First of all,this paper combines the principle of the magnetometer to measure the geomagnetism,analyzes the cause of the error of the magnetometer,and establishes the corresponding comprehensive error model.Then,for the problem that the magnetometer data fluctuates greatly and the gross difference affects the compensation result,a weight matrix is introduced to improve on the basis of the existing ellipsoid fitting algorithm,and the error compensation of the gross difference is weakened through multiple iteration calculations Impact.Then designed a physical experiment.The experimental results show that the distribution of the magnetic field strength modulus using the improved ellipsoid fitting algorithm for error compensation is more stable,the error is smaller,and it has a better compensation effect.Secondly,this paper proposes an improved square root unscented Kalman filter algorithm.This algorithm lists part of the system noise as state variables through the extended state vector method,and then calculates the noise sigma points for UT transformation,thereby reducing the impact of noise on state estimation An experiment was designed to compare the performance of the unscented Kalman filter and the improved algorithm.The experimental results show that the improved algorithm effectively reduces the impact of noise on data fusion.Then,this paper proposes a new method for pose calculation,namely,covariance cross fusion AR Kalman filter algorithm(ARKF-CI).First,the AR model matrix is introduced to characterize the state transition matrix in the Kalman filtering algorithm,the attitude of the gyroscope and accelerometer data is solved by the AR Kalman algorithm,and then the output of the AR Kalman algorithm and the covariance cross fusion algorithm are used.The results of the magnetometer's attitude solution are fused.The algorithm avoids the linearization of the nonlinear system,improves the state estimation accuracy,and at the same time determines the weight of the data source during fusion based on the covariance of the data,and improves the anti-interference ability of the algorithm.Finally,using the MPU6050 inertial measurement unit and RM3100 magnetic sensor to build the hardware platform of the heading and attitude system,a magnetic interference experiment was designed to verify the attitude resolution accuracy of the ARKF-CI algorithm.In this paper,the laboratory turntable and the established attitude and attitude system are used for data collection.The accuracy of the algorithm in this paper is evaluated by comparing the results of the attitude solution between the UKF algorithm and the algorithm in this paper.Taking the output value of the high-precision attitude and attitude system as the reference true value,the error statistics of the attitude solution results of the ARKF-CI algorithm,the improved SRUKF algorithm and UKF algorithm in this paper are obtained.The results show that,compared with the UKF algorithm,the accuracy of the heading angle calculation result of the ARKF-CI algorithm in the magnetic interference environment is improved by 19%,which effectively improves the attitude calculation accuracy of the attitude and attitude system.