Key Technologies For SINS/GNSS Tightly Coupled Integrated Navigation System

Both the inertial navigation system(INS)and the global navigation satellite system(GNSS)play important roles in the fields of economy and military.GNSS has the advantages that positioning error does not accumulate with time,and it can provide allweather accurate information of position and time.However,it is vulnerable to external signal interference.INS can provide short-term and high-precision navigation information of attitude,velocity and position.The errors of the navigation information accumulate with time,and the high-precision inertial measurement units(IMUs)are required for long-term and reliable positioning.There is a complementation between the characteristics of INS and GNSS.The integrated system can make up for shortcomings of the stand-alone system,and provides long-term and precise information of attitude,velocity,position and time.The integrated system of INS and GNSS can be classified into three constructions:loosely coupled system,tightly coupled system and deeply coupled system.Loosely coupled integrated systems utilize the position and velocity output by GNSS receivers as the observations,and correct the errors of INS with the differences of the observation and the corresponding output of INS.It is the easiest system to integrate,but can not adapt to the situation of high mobility.Pseudo-range,pseudo-range-rate and carrier phase are used as the observations in the tightly coupled system which are output by GNSS receiver,and the response speed of the system can cope with the high mobility situation.The in-phase signal and orthogonal signal in the correlated channels of GNSS receiver are used as the observation in the deeply coupled system,and the structure of this system is complex.Deeply coupled system requires high-quality and precise inertial measurement units.In summary,the tightly coupled integrated system is a relatively balanced system in all aspects.This paper focuses on the related technologies of INS/GNSS tightly coupled integrated navigation systems.The denoising of the inertial measurement units,the integrity and the availability of the tightly coupled system and GNSS ambiguity resolution aided by INS are researched in this paper.The contents are as follows:(1)In the case of the loss of satellite signal,the navigation accuracy of tightly coupled integrated system depends entirely on the inertial navigation system.Therefore,it is very important to improve the precision of INS.The error of inertial measurement units is the main factor affecting the precision of the inertial navigation system.Variational mode decomposition method is introduced to analyze the performance of the IMUs for the first time.The signals of the IMUs are decomposed into the modes with different center frequencies,and every mode is processed according to the criteria of continuous mean square error.The white noise in the modes are reduced and the signal-to-noise ratio of the signals from IMUs is improved.Then,the new signals are synthesized from each mode,and are put into the inertial navigation computer to improve the precision of the inertial navigation system.It can be proved that the signals after variational mode decomposition method can improve the precision of the inertial navigation system by comparing the static and dynamic test results of the inertial navigation which use the signals before and after variational mode decomposition.(2)In the navigation process,the tightly coupled integrated system will inevitably receive faulty observations,so detection and identification algorithm for the faulty observations is the essential requirement for the normal operation of the tightly integrated system.In this paper,a fault detection and identification algorithm based on the variance shift outlier model is studied.By establishing the restricted maximum likelihood ratio test and score test statistics of t-distribution function,the observations are tested,and the estimated variance inflated factor is brought into the covariance matrix of innovation in Kalman filter to down-weight faulty observations in the estimation and keep the stability of the system.Then,the detection and identification algorithm based on Chi-square test and Kalman filter for faulty observations of relative positioning tightly coupled navigation system is studied.Whether the observations are faulty is decided through checking the sum of squares of the innovation of Kalman filter having a Chi-square distribution.And then identify and eliminate the fault observation by w-test to avoid the contamination to the estimation.Fault detection and identification algorithms suffer from the masking effects and swamping effects.Therefore,it is necessary to compare the horizontal error with the positioning horizontal protection distance.The positioning horizontal protection distance with the horizontal alarm threshold ensures the integrity of the tightly integrated system and the effectiveness of the navigation results.It can be proved that the algorithm based on VSOM could detect and identify the faulty observations and down-weight them in Kalman filter to keep the system stable;the algorithm based on chi-square test and wtest can detect and identify the faulty observation correctly,and exclude the faulty observations from the filter to ensure the stability of the system.(3)To obtain the high-precision positioning in the relative positioning tightly coupled integrated navigation system,the ambiguities in the double-difference carrier phases must be fixed.The fixed efficiency of ambiguity is affected by the observation precision and the environment under the dynamic condition.The signals from the navigation satellites are easily affected by the complex environment in the city during the navigation process.It is difficult for some ambiguities to fix,and affects the fixing of the other ambiguities.The success rate of fixing all the ambiguities becomes low.The inertial navigation system is employed to assist the satellite navigation system in this paper since the inertial navigation system can provide the high-precision navigation information in the short term,and the Fast Ambiguity Search Filter(FASF)algorithm is applied to search and fix the ambiguities.And the strategy of partial ambiguities resolution is used instead of full ambiguities resolution.More positioning is available at the expense of reduced accuracy,and meet the requirements.The key of partial ambiguity resolution is the selection of the appropriate ambiguity subsets.The elevation angle model is used in this paper as the criterion to select the subset of the ambiguities.The experiment proves that the success rate of fixing the ambiguities is improved effectively.