Study On Kalman Filter Techniques Of Initial Alignment Of Strapdown Inertial Navigation System

Strap-down inertial navigation system(SINS)is a very important navigation device,which is mounted in automobiles,ships and aircraft and other vehicles to achieve the unmanned automatic navigation,either in the military or civilian fields have been widely application.Initial alignment is one of the basic problems in Inertial navigation systems that must be addressed,time and accuracy of initial alignment directly affects the speed of response and accuracy of SINS.How to resolve the relationship between them is a hot spot for many scholars discuss in recent years.Currently,the kalman filter plays an important role in the study of the initial alignment.It can effectively solve the problem of the initial estimate of navigation parameters,and correct components' random error.This paper around the stationary base and dynamic base to explore the initial alignment problems.Firstly,we use the cubature kalman filter(CKF)instead of the kalman filter(KF)in the stationary-base alignment.A detailed analysis of the model controllability and observability has been done.Then,establish a dimensionality-reduction CKF,the result show that the dimensionality reduction CKF is better than traditional CKF.Secondly,in the dynamic-base alignment,we use the transfer alignment method.The different calculation parameters and measurement parameters as a system view on the performance of the initial alignment measurements is summarized and analyzed.Finally,complexing the advantages of calculation parameters and measurement parameters,a system view containing the speed,velocity and attitude is established,as a combination of multiple parameter measurements in the kalman filter model.By comparing with other models,the result is show that the combination of matching algorithm is better than the other matching algorithm.Finally,considering the initial alignment problems under non-ideal filter model,based on the square-root cubature kalman filter(SCKF)algorithm and Sage-Husa algorithm,an adaptive algorithm follows the current matrix of the measurement noise which is updated by current Covariance matrix is proposed.Then,we use it in the combination of multiple parameter kalman filter model and draw a conclusion through comparing others.The simulation experiments show that the adaptive algorithm can improve the misalignment angle error estimation results.