Research On Initial Alignment And Navigation Calibration Of Strapdown Inertial Navigation System

Whether in the military or civilian fields,the status of navigation is increasing day by day,and people's reliance on navigation is increasing.Among them,the Strapdown Inertial Navigation System(SINS),due to its simple structure and easy integration of equipment,has developed rapidly as soon as it is proposed,and has been applied in many fields.SINS work can be divided into three main steps: sensor data calibration,initial alignment,and inertial navigation solution update.Among them,the initial alignment is a very important part,and the accuracy of the alignment result directly affects the final navigation result.At present,the alignment research under the static base is quite mature,and the accuracy can fully meet the needs.However,in reality,there are very few absolutely static environments,and most scenes will have shaking interference.This paper focuses on this problem,mainly studying the initial alignment of SINS under sloshing interference,and then doing calibration research on the defects of SINS that are prone to divergence.The specific content is as follows:First,under sloshing interference,due to the influence of external interference angular velocity and linear velocity,the analytical coarse alignment can no longer work normally.In this paper,an indirect alignment algorithm based on the inertial system is introduced,which uses the cone of gravity acceleration in the inertial system.The movement phenomenon,through the conversion relationship between the coordinate systems,reduces the influence of the interference angular velocity,and completes the initial coarse alignment of the SINS under the shaking base,paving the way for the subsequent fine alignment.Second,in the fine alignment stage,in order to solve the problem of filter divergence caused by inaccurate modeling and noise parameter estimation,this paper proposes an improved Adaptive Kalman Filter(IAKF)algorithm.The algorithm first draws on the idea of fault detection and isolation,and uses the sequence inertia filtering method to update the measurement noise.In the update process,different measurement noise update methods are selected according to the filtering evaluation result,which improves the reliability of the filtering.Secondly,I learned the idea of fading factor filtering,and added the control function to the gain matrix calculation of the traditional Kalman filter to speed up the convergence speed.Third,because SINS needs to perform a large number of integral operations,long-term autonomous work is prone to divergence.To solve this problem,this paper proposes a calibration method combined with speech recognition technology.This method uses the position-coordinate index table to input people through voice The geographic indications are converted into actual latitude and longitude information,which is used to correct the deviation of the inertial navigation.It greatly simplifies the calibration process,and is less dependent on external auxiliary information,which is convenient for use in some special occasions.In the end,the proposed algorithm was verified by simulation and actual measurement.Simulation experiments show that the improved algorithm can converge to the true value under different interference conditions,and the pitch angle,roll angle and heading angle errors can all be stabilized at about 10 divisions,with high stability.The experiment of using Redmi K30 Pro to collect data with shaking interference shows that the improved algorithm has a higher convergence consistency.In the same azimuth,data collected multiple times,the heading angle difference is within 5 degrees,and the pitch and roll angles are different.Within 10 divisions;while the heading angle difference of the traditional method is more than 20 degrees,the improved method is more stable than the traditional method,and the variance of the alignment result is smaller,which is more conducive to practical application.In addition,the calibration method was verified by simulation,and the feasibility of the method was discussed.The first version of the calibration platform model was made in conjunction with the low-power device STM32,and the actual calibration method was tested to verify the validity of the calibration method.