| In recent years,with the continuous increase in the research heat of unmanned vehicles,Mars probes and manned submersibles,navigation systems have become more and more important,and MEMS magnetometers,MEMS gyros and fiber optic gyroscopes are used as navigation systems.The important attitude information sensor,its environmental adaptability and anti-interference ability can directly affect the final navigation accuracy.Based on this,this paper focuses on the environmental magnetic interference error,random noise error and temperature drift error generation mechanism and its suppression methods,including the following aspects.In this paper,the basic principles of MEMS magnetometer,MEMS gyro and fiber optic gyroscope and the structural components of each link are introduced in detail.The specific models and main functional indexes of the sensor are described in this paper.The error generation mechanism and basic characteristics are analyzed in detail.Expand to lay the foundation of the model.Based on the environmental magnetic interference compensation and the coupling effect between soft and hard magnetic coefficients,and considering the excessive difference of the order of magnitude of soft and hard magnetic coefficients,a model of least squares two-step lag compensation algorithm based on Newton-Raphson method is established.That is,segmentation optimization is used in the compensation of hard magnetic and soft magnetic coefficients,and the second optimization is performed by Newton method.The validity of the model is verified by the simulation and laboratory test based on the international geomagnetic reference field IGRF-12.The error estimate is reduced from 30?T to 0.7?T.This model can effectively improve the navigation output accuracy and reduce the error of the total magnetic field strength.The random noise error caused by various disturbance moments caused by environmental factors on MEMS gyroscope is studied in detail.Firstly,according to the function index of MEMS gyroscope and the characteristics of random noise,the mathematical model of random noise error is established.The random noise error in the gyro output signal is denoised mainly by EMD-SVD method,EMD-LW method,Kalman filtering method based on ARMA model and Kalman filtering method based on GA optimized PW method.Then the experimental test of the MEMS gyro output signal is carried out,and the denoising effect of the above algorithm on random noise error is discussed respectively.The experimental results show that the Kalman filtering method based on GA optimized PW method has the best effect in the error compensation of gyro random drift.,can effectively improve the signal to noise ratio and reduce the mean square error.In addition to MEMS gyros,the development of fiber optic gyroscopes has also attracted attention in recent years.Therefore,this paper studies in detail the temperature drift error caused by the thermo-induced non-reciprocal effect caused by ambient temperature changes on fiber optic gyroscopes,first based on the heat of fiber optics.Based on the distribution of light effect and temperature variation,the mathematical model of temperature drift error of fiber optic gyroscope in the temperature range of(-20°~50°)was established.Polynomial piecewise fitting,over-limit learning machine neural network segmentation compensation algorithm and Elman were respectively used.The neural network segmentation compensation algorithm performs temperature drift error compensation,and experimentally tests the fiber optic gyroscope output signal under variable temperature conditions.The suppression effect of the above algorithm on temperature drift error is analyzed and analyzed.The experimental results show that the latter two algorithms have temperature drift.The suppression effect of the error is obviously better than the first algorithm.The Elman neural network segmentation compensation algorithm can make the fiber optic gyroscope maintain better zero-bias stability when the temperature changes. |
PDF Full Text Request