Showing papers in "Journal of Chinese Inertial Technology in 2012"

Inertial navigation algorithm aided by motion constraints of vehicle

Abstract: In order to improve the independent navigation accuracy of vehicular SINS and enhance the mobility of the car,the integrated navigation method aided by the motion constraints of the vehicle was adopted.By making use of the condition that the cross-track and vertical velocity of the normal running vehicle were zero,and taking the installation errors of SINS into consideration,the Kalman filter algorithm was deduced,the vehicular navigation experiment for about 80 km was implemented.The results show that the level positioning precision was 16.6 m(CEP) and the altitude precision was 14.9 m(PE).This method can be used as a backup navigation strategy to be used when the odometer was invalidated,and can also be used to extend the interval time of stopping car in ZUPT(zero-velocity update) mode.

Calibration of gyro error model coefficients on precision centrifuge with counter-rotating platform

Abstract: In order to accurately calibrate the gyro drift coefficients,a coordinate system was set up based on the precision centrifuge with counter-rotating platform,and the attitude relationship between two adjacent coordinate systems were analyzed in consideration of the error sources.Precise angular velocity input for each axis of gyro was derived by using direction cosine matrices,and the nominal value of each specific force input was given as well.Combined with the gyro’s static error model,various harmonic amplitude expressions were deduced including gyro drift coefficients.By using centrifuge to provide two different harmonic accelerations,and sampling the output of gyro at each sampling time,the amplitudes of 0-2nd harmonics were obtained through Fourier analysis.The drift coefficients were identified by least square method based on the relationship between harmonic amplitude expressions and drift coefficients.The test simulation results show that the calibration of the gyro drift coefficients on centrifuge with counter-rotating platform can effectively avoid some of the centrifuge errors,and other centrifuge errors only have significant impacts on DI,DS,DIO and DOSwhose accurate calibration can be achieved by compensating the centrifuge errors.

INS in-motion alignment for land-vehicle aided by odometer

Abstract: In-motion alignment plays an important role in improving the maneuvering capability of the carrier,which allows the initialization of an inertial navigation system(INS) while on moving.Different from stationary alignment,the implementation of in-motion alignment usually needs an external navigation device such as GPS or odometer to provide vehicle dynamic information to compensate and correct the INS errors.In the field of land navigation,odometer aided in-motion alignment is widely adopted due to its full self-contained characteristic.In this contribution,odometer error is modeled comprehensively as a part of system dynamic model,and the displacement incremental matching method is proposed to finally realize the integration of odometer and INS.Besides,effective fault diagnosis is proposed to evade odometer measurement errors caused by the complicated environment.The effectiveness of the proposed method is verified through ground based alignment experiments.The results show that the heading error is less than 0.05° after ten minutes initial alignment which is equivalent to that of the stationary alignment.

Initial alignment algorithm for SINS based on quaternion Kalman filter

Abstract: An anti-disturbance initial alignment algorithm for swing-base strapdown inertial navigation system(SINS) is presented,and the essential of the proposed approach is attitude determination(AD) using vector observations.It is easy to show that the apparent gravity expressed within an Inertial Space forms a cone whose main axis is the rotational axis of the earth.So,by the examination of gravity movement,the true heading can be found.Employing the gravity vector projective relation with the initial attitude-quaternion,the linear pseudo-measurement equation is built up.The initial attitude-quaternion is a constant parameter and the quaternion kinematics equation is linear and precise.The above characteristics yield a linear KF that eliminates the usual linearization procedure and is less sensitive to initial estimation errors.The turntable rocking simulation experiments are carried out which demonstrates the efficiency of the quaternion filter.

Effect of octupole-winding on temperature performance of fiber optic gyroscope

Abstract: By studying the octupole-winding method,the improvement of the fiber optic gyroscope’s(FOG’s) temperature performance was realized.The element model of the fiber coil with the octupole-winding pattern was built based on the discrete mathematics formulae of Shupe error in FOG.A turn-by-turn quantization Shupe bias error model was established in octupole-winding method.Based on the temperature distribution model mentioned above,the effects of octupole-winding coil and quadrupole-winding coil on the temperature performance of FOG were simulated under the same radial temperature gradient.The results show that the octuple-winding method could reduce the Shupe bias errors by 80% in comparison with the quadrupe-winding method,and this agrees with the results of the experiments.

Personal navigation method based on foot-mounted MEMS inertial/magnetic measurement unit

Abstract: In order to relieve the dependence of personal navigation on global navigation satellite system(GNSS),a personal navigation method was studied based on foot-mounted MEMS inertial/ magnetic measurement unit.Navigation information including attitude,velocity and position was obtained through the data from MEMS inertial measurement unit and strapdown inertial navigation algorithm.The azimuth of the navigation system was obtained by magnetic sensors.The errors of MEMS inertial navigation system and random errors of inertial sensors were modified by applying gait phase detection and zero-velocity update,so that the accumulate speed of the positioning errors was slowed down.The navigation experiment result shows that,the navigation errors of straight line and rectangle route retain about 2 m and 6 m during about 9 minutes’ walking.The errors take up 1.1% and 2.5% of the whole distances of walking respectively.The experiment conclusion proves that the proposed method can effectively improve the positioning precision of personal navigation system,and can achieve personal positioning for longer time under the environment that GNSS signal is attenuated or become failure.

Integrated navigation method aided by skylight polarization

Abstract: In view of the limitations of conventional INS/GNSS integrated navigation method,an orienting technique based on polarized light was studied.First,the atmospheric polarization mode in clear sky was explored,and the ideal atmospheric polarization characteristic was revealed.Second,the principle of polarization navigation sensor was analyzed.In order to change the orientation value measured by polarization sensor to the heading,the calculation equations to determine solar position was derived.At last,the structure of SINS/GPS/Polarized-Light integrated navigation system was designed,and the dynamic simulation was carried out.The simulation results indicate that the system's heading error can be corrected by information from polarization sensor effectively,and the heading accuracy is equivalent to that of polarization sensor.The independent orienting technique based on polarized light can enhance the system's navigation performance and has wide application prospect.

Modeling and simulation of transfer alignment for distributed POS

Abstract: To improve the precision of slave POS(position and orientation system) in distributed POS and compensate for the wing's flexure deformation of aircraft,a novel transfer alignment method was proposed.The ground-demo of the distributed POS was set up to test the proposed method.A mechanical method with the aid of the ANSYS software was used to establish the motion equation of the model wing.Then the flexure angle and its angular rate were added to the state variables of the Kalman filter.Based on this,the attitude error measurement equation was deduced with considering the flexure.The Kalman filter using the "velocity+attitude" matching algorithm was designed.The simulation results show that the precisions of horizontal angle and azimuth angle are superior to and respectively by the proposed method.All these characteristics show that the scheme is feasible and can provide references for the engineering application of the distributed POS.

Discussions on observability and its applications in SINS

Abstract: The degree of observability was proposed to analyze the observable degree of the combined states of linear systems.It was applied in the analysis of the efficacy of the Kalman filter.Two existed methods for calculating the degree of observability were introduced,and the equivalence of the two methods was proved.The deficiencies of the existed methods were analyzed from a theoretical perspective,and a typical example was given to verify the theoretical results.A new method to determine the degree of observability of the combined states was given which was more comprehensive than that of the existed ones.In the end,the new method was applied to the observability analysis of strapdown inertial navigation systems.The results demonstrate that the new method can quantify the observable degree of the states of the system and is a great complement to the analytic method.

Variable parameter gyrocompass alignment algorithm based on large error angle model

Abstract: To solve the initial self-alignment problem of strapdown inertial navigation system(SINS) under large error angle condition,a new gyrocompass alignment scheme is put forward by analyzing SINS large error angle model and based on the principle of compass effect in platform inertial navigation system(PINS).The scheme can be divided into three steps: horizontal alignment with azimuth angle uncertainty;time-vary parameter gyrocompass alignment for large error angle;fixed parameter gyrocompass alignment.Converge time can be shorten by adjusting gyrocompass parameter in real time in this scheme.A more accurate error transfer mode in strapdown inertial navigation system(SINS) is depicted by large azimuth error angle model instead of small error angle model.At last,a simulation is made using this scheme which shows that the gyro bias is stabilized at 0.01(°)/h,and azimuth precision can reach in 60 s and at end of alignment.

Transfer alignment of carrier-born aircraft under large azimuth misalignment angle

Abstract: To improve the carrier-born aircraft’s viability and the shooting percentage of airborne missile,the initial alignment of INS in carrier-born aircraft must be finished accurately and quickly.The transfer alignment is always used to solve the problem of initial alignment due to the influence of complex marine environment and many interference factors.In this paper,the transfer alignment of large azimuth misalignment was studied,and the nonlinear error model of velocity plus attitude matching was established.Then the transfer alignment algorithm of large azimuth misalignment based on the UKF was designed.The simulation results show that the maneuver manners make no difference on transfer alignment of carrier-born aircraft in moderate sea state,and by using different filter frequencies of UKF,the transfer misalignment error at 20 Hz is about half of that at a lower frequency.

FOG random drift modeling by artificial fish swarm algorithm

Abstract: Using historical data to accurately set up FOG random drift model plays an important role in improving the precision of FOG strapdown inertial navigation system.This paper details the artificial fish swarm algorithm(AFSA) and improved artificial fish swarm algorithm(IAFSA),and gave the steps and methods for random signal modeling by AFSA.Models of FOG random drifts are set up by the traditional time series analysis,AFSA and IAFSA,respectively.The modeling results show that,the FOG Random Drift Model by AFSA is accurate and its precision is 1.5% higher than that of the traditional time series analysis.AFSA and IAFSA are both superior to the traditional time series analysis method in complexity or precision of modeling.The model by IAFSA is more accurate than AFSA,and its convergence speed is faster than that by AFSA.

Ultra-low speed detection method for CMG gimbal servo systems

Abstract: At ultra-low speed,the gimbal speed resolution is limited,and this could significantly influence the speed control of control moment gyro(CMG) gimbal servo system.Based on the analysis about the influence,a multicycle backward speed difference method is proposed to increase the gimbal speed resolution and maintain the speed bandwidth.But the detection phase delay produced by multicycle backward difference restricts further improvement of speed resolution.Therefore,based on proper difference cycle,a Kalman estimation is introduced to further increase the gimbal speed resolution and the bandwidth.Simulation results demonstrate that the speed resolution is increased by two orders of magnitude by the proposed methods,and the creeping phenomenon of gimbal servo system caused by limited gimbal speed resolution is avoided.

Storage life and reliability evaluation of accelerometer by step stress accelerated degradation testing

Abstract: The method of step-stress accelerated degradation testing(SSADT) is utilized to evaluate the storage reliability and life of accelerometer.First,based on failure mode effect and criticality analysis(FMECA) as well as fault tree analysis(FTA),it is determined that temperature stress is its sensitive stress,and Arrhenius model is taken as its accelerated model.Then,an appropriate liner model was built with the test data.Combined with the threshold of failure,the storage life of each sample can be calculated.The hypothesis test show that the storage life complies with Weibull distribution pattern.The distribution parameters of the accelerometer under normal temperature are obtained with the accelerated model.The reliability of the accelerometer is 0.988 when it has stored for 140 160 h(16 a).

Multiple fading factors Kalman filter and its application in SINS initial alignment

Abstract: In view of poor robustness of Kalman filter algorithm when there exits model errors or unknown external disturbances,a multiple fading factors Kalman filter is proposed for a kind of linear model,which is based on the property that the sequence of residuals is uncorrelated when Kalman filter gain is optimal.The new algorithm can adjust multiple fading factors adaptively,therefore it can fade multiple data channels.When the filter becomes steady,it can also adjust the filter gain,which makes it have strong robustness when there exits model errors or unknown external disturbances.The proposed algorithm is applied to strapdown inertial navigation system(SINS) initial alignment with inaccuracy noise statistics,and simulation indicates that,compared with those of single fading factor Kalman filter and conventional Kalman filter when system model is inaccuracy,the estimating accuracy of heading error angle of the proposed algorithm is improved by 70% and 43% respectively,which proves the effectiveness of the new algorithm.

Three-dimensional magnetic compass error compensation algorithm based on ellipsoid surface fitting

Abstract: To solve the problems in existing 3D magnetic compasses(such as,error compensation is slow,special turntable is needed,and field applying is inconvenient),we put forward a method to establish a scalene ellipsoid surface mathematical model of the general magnetic field to compensate errors.First,we established the ellipsoid surface model according to the general magnetic field model presented by Poisson equation.Then,we linealized the ellipsoid surface equation using Newton iteration method,and used least square method and total least square method respectively to calculate linear equations,derive the mathematical forum of ellipsoid surface fitting and the ellipsoid surface parameters.Finally,we derived the calculation forum of parameters’ initial value by using least square method,and gave the method of selecting data point.The results show that the errors can be compensated only by using a hand to hold a magnetic compass and let it spin in all quadrants.And the heading measurement precision can reach 0.8°.The compensation precision of this method is almost the same as that of traditional turntable compensation methods,but it is more agile and fit for the field use with no turntable equipment.

Design of horizontal damping network for INS based on complementary filtering

Abstract: Schuler oscillation damping is one of the key technologies to improve the long-term precision of INSIt is difficult for inertial systems using low-order networks to perfectly perform restraining on both high-frequency and low-frequency reference velocity errorsIn this paper,a high-order horizontal damping network based on complementary filtering was designed,in which the damped Schuler loop may obtain 40 dB/10 deg or higher attenuation to both high-frequency and low-frequency reference velocity errorsSuch single loop is equivalent to two-loop of integrated system using low-order damping network which perform good in high frequency and low frequency respectivelyComputer simulation and sea trial results show that inertial systems adopting the designed high-order network achieve excellent filtering characteristics in both high frequency and low frequency,which are better than those using low-order networks and have merits in engineering applications

Mixed linear regression temperature compensation method for annular-vibrating MEMS gyroscope

Abstract: Annular-vibrating MEMS gyroscope is significantly affected by temperature and it also has strong self-return characteristics.In view of the problem that traditional method(such as sectional fitting) could not accurately compensate the gyroscope influenced by temperature,a mixed linear regression model is proposed to set up a temperature compensation model.According to the characteristics of mixed linear regression model,this method takes the influences of gyroscope itself and the temperature changes as well as other factors into account.The multiple linear regression method is also used in determining the coefficients.At last,normality test for the residual is used to check if the model can meet the change rule of gyro data.Experimental results show that the mean of the gyroscope can be reduced by one to two orders of magnitude.The compensation method has a good reproducibility,so it has a great value of engineering reference.

Relative navigation between vessels based on GPS single difference

Abstract: In order to get relative position information during flotilla navigation,GPS pseudo-range information was observed and exchanged between two vessels via flotilla communication link to get single difference.The single difference was used to establish the relative navigation model between two vessels as observation measurement.The relative position information such as relative distance and bearings was taken as state measurements,and the least multiply square algorithm was applied to resolve the relative vector between the two vessels.The relative distance and bearings between two vessels in WGS84 coordinate frame system is calculated,and experiments were carried out.The data processing results show that the relative distance determination accuracy is prior to 0.2 m,and the bearings determination accuracy is prior to 5°.The results verify that the relative navigation model between two vessels is feasible,and the accuracy reaches the performance requirement of relative navigation.

SINS/CNS nonlinear tightly integrated method based on additive quaternion

Abstract: A SINS/CNS nonlinear tightly integrated method based on additive quaternion was put forward in accordance with the nonlinear characteristics of SINS/CNS integrated navigation system.From the nonlinear error modeling of SINS and the nonlinear measuring of CNS,the error transmission property of strapdown inertial navigation based on additive quaternion was deduced,and a nonlinear measurement model for celestial navigation was constructed.SINS and CNS nonlinear information fusion was also realized using 2-order interpolation filtering algorithm.Computer simulations demonstrate that the SINS/CNS nonlinear tight coupling method takes the system’s nonlinear characteristics into account,and its peak error in maneuvering flight situation is small.The navigation precision by this method can increase by about 15% compared with that by linear tight integrated method.

Analysis and simulation on anti-multipath performance of Beidou2 navigation

Abstract: An evaluation method of code tracking multipath error envelope based on coherent Early minus late delay lock loop is derived to analyze the anti-multipath performance of Beidou2 navigation signal.The simulation shows that all Beidou2 navigation signals are able to discriminate the multipath delay more than 300 m under conditions of multipath-to-direct ratio(MDR)=-6 dB,correlation distance d=24 ns and receiving bandwith =24 MHz.The anti-multipath performance of navigation signal is simulated under different multipath-to-direct ratio(MDR),correlation distance,signal received bandwidth and signal composition.The results show when MDR is a constant,the signal multipath error will be limited to 4 m and decline fast if increasing the high frequency component of Beidou2 navigation signal and utilizing narrow correlation distance and wide receiving bandwidth.So the signal's anti-multipath performance can be improved.

Effect of thermoelectric coolers on temperature control of SLD

Abstract: Aiming at the poor matching between SLD and its control circuits,this paper analyzes the effects of SLD's semiconductor cooler on discretization of temperature control parameters and gives the solving methods.Results indicate that: 1) the main reasons of parameter discretization are the characteristics of equivalent resistance and cooling efficiency of TEC;2) Under the present temperature control mode,the equivalent resistance of TEC is the most important reason of parameter discretization,depending on techniques and conditions;3) The cooling and heating efficiency of TEC may accelerate or prevent dispersing of temperature control parameters.Changing the mode of temperature control,building the standard management of SLD,improving the consistency of TEC,and increasing the cooling efficiency of TEC would improve the match effects between SLD and its control circuits.

Analysis and compensation for size effect of strapdown inertial navigation system

Abstract: The size effect error is the most importation error source of strapdown inertial navigation system(SINS)In this paper,the mechanism and effect on the system precision of the size effect error were analyzed in theoryBased on this,a two-sample compensation algorithm for the size effect error was deduced by using polynomial approximationBesides,the rate calibration method of size effect error parameters was proposed according to the reason and regularity of the size effect errorTheoretical analysis and experiment results show that the size effect error is the system's inherent error formed by its physical structure when with angular motion,and its compensation should adopt both hardware design and software algorithmBy using the two-sample compensation algorithm,the accelerometer's size effect error is reduced from 006g to 1×10-4g,and the horizontal positioning error is reduced by about 10℅

Time synchronization on SINS/GPS integrated navigation system under different dynamic conditions

Abstract: SINS/GPS integrated navigation systems which have different accuracies and motion states have also different requirements for the time synchronization errors.However,most of the time synchronization solutions adopt a uniform method for compensation,not for specific application objects.It could lead to issues of the high cost of hardware or complex software algorithms.In order to find the most suitable and the lowest-cost time synchronization scheme,this paper analyzes the affection of the time synchronization error for system accuracies,and then the appropriate hardware or software synchronization solutions are proposed in the low,medium and high dynamic situations which are method 1,methods 2 and method 3 respectively.Through vehicle tests and software simulations for each scheme,the results verify that it is not obvious for improving the system accuracy by applying three methods in the low dynamic situation,and under the medium dynamic condition,the accuracies of the attitude angles and velocities for method 2 and method 3 are improved about 50%,while the position accuracy is improved about 20% than that of method 1.Method 3 shows the most significant effects in the high dynamic situation because the accuracies are all improved by more than 70% than those of method 1.

Approach for detection of slowly growing errors in INS/GNSS tightly-coupled system based on LS-SVM

Abstract: Aiming at improving the instantaneity of detecting slowly growing errors with the autonomous integrity monitored extrapolation(AIME) method in the INS/GNSS tightly-coupled navigation system,this paper presented a new method based on least squares support vector machine(LS-SVM) combined with AIME.The method replaces the innovation of Kalman filter by the forecasted innovation of the Least squares support vector machine regression,which is used for the test statistics.The new method presented in this paper could reduce the influence of fault tracking caused by Kalman filter,so the real-time ability of slowly growing errors detection can be improved.Simulation results show that,the delay time of slowly growing errors detection is reduced by 1.3 times.The method based on Least squares support vector machine has better detection ability than traditional autonomous integrity monitored extrapolation method when detecting slowly growing errors.

Modeling and analysis of fiber optic gyroscope random drifts

Abstract: The precision level of fiber optic gyro is the key factor of inertial navigation system,and reducing the gyro random drifts is an important means to improve the precision.Building the model of gyroscope random drift and filtering it can effectively improve the output precision of fiber optic gyro and then improve the accuracy of inertial navigation system.In this paper,the ARAM model of fiber optic gyro random drifts was built through numerous experiments,in which the random drifts were effectively filtered.The filtering results were analyzed by Allan variance analysis.The result shows that the main error sources and the sinusoidal noise in fiber optic gyro output signal are reduced to about 50% of that before filtering,and the high frequency noise is effectively restrained.It is proved that the fiber gyro random drift model is effective and feasible.The visualization software was designed at the same time,makeing it valuable for engineering application.

Analysis and design of closed-loop control system for silicon resonant accelerometer

Abstract: The silicon resonant accelerometer(SRA) has become a hot spot recently,due to its taking quasi-digital signal as output and the potential of high precision.The closed-loop control is one of the important ways to improve its performance.Based on the analysis of the characteristics of structure and circuits of the SRA,two schemes of magnitude auto gain control(AGC) were compared,and two kinds of phase closed-loop schemes of phase-shift and phase lock loop(PLL) were analyzed.The AC AGC and PLL were chosen as the final closed-loop system scheme of SRA through experiments.The natural frequency of the designed accelerometer is 17 kHz with a scale factor of 220 Hz/g,and the bias stability of SRA is 21.9 μg in 2 h.

Automatic landing system based on INS/ILS/RA integrated navigation

Abstract: During precision approach and landing,the civil aircraft is susceptible to the receiver noises and space noises such as field environment and electromagnetic interference.In view of the problems in filtering out the phase lag of beam error signal,an automatic landing system(ALS) based on INS/ILS/RA integrated navigation is put forward.The measurement signal is an error signal achieved by comparing the output position of the inertial navigation system(INS) with that of the instrument landing system(ILS) and radio altimeter(RA).Then using a Kalman filter to estimate the results,the beam deviation will be replaced by two outputs of the integrated system-the course angle and pitch angle,which are then looped back to the control circuit respectively.The simulation results show that the directions angle error is superior to 0.3°,and the gliding angle error is superior to 0.2°.The integrated system can significantly improve the dynamic quality of the beam error control signal,reduce the noise influence and improve the stability and closed-loop performance.

Fixed-interval smoother and its applications in integrated navigation system

Abstract: As a post-processing estimation method,smoother is superior to optimal state estimation algorithm for the integrated navigation systems by off-line processing or with low real-time requirement,since the smoother uses more measurements than the filter.Fix-interval smoothing estimates the state after processing all of the measurements in the interval,and it has been widely used in the integrated navigation systems.The Rauch-Tung-Striebel smoother and the two-filter smoother for the linear systems are mainly expounded,and the nonlinear smoothing methods based on the formulations of those two smoothers are also introduced in this review.Moreover,the advantages and disadvantages as well as the applications of these smoothing methods are summarized.

Parameter calibration based on attitude errors of space-stable INS

Abstract: In order to improve the attitude accuracy of space-stable INS,the navigation attitude error was used to calibrate the systematic parameters.First,the error model of attitude errors was given considering the gyro drift,accelerometer error,misalignment angle of gyro case rotation axis,installation error and frame angle bias.Then,the separable parameters and calibration method were summarized up,and based on this,the experiment scheme was designed to identify the systematic parameters.At last,the attitude errors obtained from the experiment were used to identify the systematic parameters by least square identification method.Experiment results show that the accuracies of roll,pitch and yaw have improved 60%,60% and 40% respectively after compensation,and the calibration errors are reduced to 15% of the attitude accuracy index.