Research On Initial Alignment And Error Suppression Technology For Shipborne Navigation System

In order to ensure safe navigation,ship always equipped with strap-down inertial navigation system(SINS),Doppler velocity log(DVL),Global Navigation Satellite System(GNSS),celestial navigation system(CNS)and other navigation systems.The integrated navigation is the most popular navigation method for ship,which combines the advantages of SINS and other assisted navigation systems.Due to the rapid start-up requirements and the long-term use of navigation sensors,the process of aging,maintenance and replacement,and even the impact of marine complex environmental interference,shipborne navigation technology still needs to be further improved in many aspects,including in-motion initial alignment technology,online calibration technology of the navigation system and the integrated navigation information fusion technology all need to be further improved.To improve the navigation accuracy of ship,the research on initial alignment and error suppression technology for shipborne navigation system are carried out in this dissertation,focusing on the research of in-motion initial alignment for SINS,online calibration and information fusion filtering algorithm for integrated navigation system.The details are as follows:1.Researches on in-motion initial alignment for shipborne SINS.Firstly,The global observability analysis results show that the lever arm and installation error of GNSS or DVL relative to SINS will be introduced into the observation vector during the maneuvering process,which may reduce the accuracy of constructed observation vector.Secondly,to reduce the effect of lever arm and installation error of GNSS or DVL,an in-motion initial alignment method based on error compensation is proposed,which may reduce the accuracy of constructed observation vector.The performance of in-motion initial alignment method based on error compensation is verified by simulation.Thirdly,To weaken the influence of unknown lever arm and installation error,a Kalman filter based in-motion initial alignment for DVL aided SINS is proposed.A closed-loop scheme is presented to simultaneously estimate and compensate the initial attitude matrix,lever arm and installation misalignment angle error by using Kalman filter,which weakens the effects of these unknown parameters and improves the accuracy of vector observations.The global observability analysis method is utilized to analysis and design a special trajectory for ship,which ensures the unknown parameters can be observed.Finally,the simulation test for in-motion alignment is carried out to verify the performance of the proposed Kalman filter based in-motion initial alignment method.2.Research on online calibration method for SINS/DVL/GNSS/CNS integrated navigation system.To improve the observability of IMU calibration error and efficiency of calibration for integrated navigation system,a new online calibration scheme for SINS/DVL/GNSS/CNS integrated navigation system is presented in this thesis.Firstly,aiming at uncertainty observability of some calibration parameters,the observability analysis is conducted to determine the observability of different online calibration parameters,and the analysis results are used to simplify the calibration state-space model and design the motion trajectory.Secondly,a information fusion strategy based on the sequential filter and a closedloop feedback algorithm is proposed to realize the information fusion of integrated navigation sensors with different output frequencies,and the degration of Kalman filter caused by nonlinear error of calibration parameters can be restrained.Finally,the simulation test is carried out to verify the validity of the proposed online calibration scheme.3.Research on SINS/DVL integrated navigation method based on Robust adaptive filtering algorithm.To suppress the effect of measurement outliers and imprecise filtering parameters,a robust adaptive filter is proposed for SINS/DVL integrated navigation system.First,the state transformation state-space model for SINS/DVL integrated navigation system is derived to avoid the problem of variance inconsistency caused by the specific force term.Second,for imprecise filtering parameters,the one-step predicted covariance matrix is modeled as IW distribution,which can weaken the effect of imprecise covariance matrix of state noise indirectly by adaptive estimation of one-step predicted covariance matrix.To detect the measurement outliers,the measurement likelihood probability density function is involved in a binary indicator variable modelled by Beta-Bernoulli hierarchical prior.Then,the navigation errors of SINS/DVL integration are estimated by the new robust adaptive Kalman filter,where the state vector,one-step predicted covariance matrix,Bernoulli variable,and Beta variable are jointly estimated through variational Bayesian(VB)method.Finally,the navigation performance of SINS/DVL integrated navigation method based on robust adaptive filtering is verified by simulation and sea trial experiments.