Design Of Integrated Navigation System Based On FOG-SINS And Star Tracker

The inertial navigation technology based on fiber optic gyroscope(FOG)has been successful in various applications due to its high reliability,small size,low energy consumption,and augmented performance.This thesis focuses on the design of an integrated navigation system based on fiber optic gyroscope strapdown inertial navigation system(FOG-SINS)fused with an accurate star tracker attitude sensor in order to obtain precise estimation of the navigation information.Firstly,a model of the inertial measurement unit(IMU)based on FOG and MEMS accelerometer is established.The sensors errors are analyzed rigorously,special interest is given to the random noises.The modeling of the noise existing in sensors has a great effect on design of the fusion filter.In order to achieve better estimation of the random process parameters,a novel stochastic modeling algorithm based on adaptive dual threshold wavelet function and Allan Variance is proposed.Experimental results based on real data from FOG-IMU shows a satisfactory improvement in denoising performance.Secondly,attitude algorithm is a substantial part of this work.Several different attitude algorithms are discussed.Rotation vector algorithm has an advantage of correcting the noncommutativity error resulting from an oscillatory rotating motion.The algorithm performance is affected by the chosen sample number.While large sample number may reduce the update rate,this may cause conflict between accuracy and complexity.In order to solve this problem,an adaptive multi-sample SINS attitude algorithm is presented.A two-speed rotation vector algorithm with a predefined threshold is applied based on estimation of the oscillating motion.Furthermore,new adaptive coning coefficients algorithm is introduced for optimizing the compensation results.Thirdly,the design of fusion filter for FOG-SINS and star tracker is discussed.The dynamic model of FOG/INS/Star Tracker integrated system is set up based on the previous work.After a performance evaluation of different nonlinear filters,an adaptive iterative modified Sage-Husa filter is nominated.The filter only requires an initial approximation,then the covariance is validated harnessing available measurements.Finally,the results of the stochastic modeling of real data from FOG-SINS using Allan variance on data denoised by an adaptive dual threshold wavelet denoising function are integrated to the fusion filter.A simulation system based on real FOG-IMU and a star tracker simulator is devised for testing and experiments.Experimental results show satisfactory performance of the presented integrated navigation system.The accuracy of the attitude,velocity,and position has been improved in comparison to standalone FOG-SINS system and higher updating rate compared to the star tracker sensor.The research achievement of the thesis is putting the foundations for further improvement of the integrated navigation solutions based on FOG-SINS and star tracker,as well as widening the application area of the sensors.