Research On GNSS/MARG Coupled Navigation Method

Low cost coupled navigation develops rapidly in navigation system,but the low-cost INS(Inertial Navigation System)can not work independently for a long time because of its low precision and large error.The positioning error of GNSS(Global Navigation Satellite System)does not accumulate with time,but the satellite signal is vulnerable to external interference.The magnetometer can provide azimuth with small error without interference and does not accumulate error.In order to effectively use low-cost sensors for high-precision navigation,this paper uses the advantages of the above three to build a multi-sensor coupled navigation system based on the MARG(magic,angular rate and gravity)and GNSS.Aiming at the problems of poor accuracy,large calculation and easy to block GNSS signal in the practical application of the coupled navigation of GNSS/MARG.In this paper,different combination methods are studied,including Kalman filter,GNSS error prediction and model switching.The main work of this paper is as follows:First,there are two kinds of combination of GNSS/MARG: position velocity combination and pseudo range pseudo range rate combination.The mathematical model of the related system is analyzed and established.First of all,the magnetometer,accelerometer and gyroscope in MARG are combined,and the information of position and speed obtained by output is combined with the position and speed obtained by GNSS;secondly,the pseudo range and pseudo range rate calculated by MARG and GNSS ephemeris are combined with the pseudo range and pseudo range rate obtained by GNSS receiver,and the error equation in coupled navigation is established,and the measurement equation of Kalman filter is constructed The system optimization comparison in the following paper is the foreshadowing.Second,in order to improve the accuracy of the pseudo range pseudo range rate combination scheme and solve the problem of large amount of calculation,the paper makes full use of the linear stability of the volume Kalman filter and the simple structure of the extended Kalman filter,and reduces the coupled navigation error after combining the two schemes.On this basis,the sequential update method is used to greatly reduce the calculation amount of the Kalman filter.The quantitative analysis and experimental verification of the filtering scheme before and after the improvement show that the calculation amount of the scheme is reduced by more than 20%,and the accuracy is improved by more than 10%,all of which have significant advantages.Third,in order to determine the combination scheme that is more suitable for the actual environment,the two combination schemes are tested in different environments.It is found that the pseudo range pseudo range rate combination scheme increases with the increase of the number of satellites,and the positioning speed slows down,while the accuracy of the speed position combination mode is close to the pseudo range pseudo range rate combination,and the calculation amount is unchanged,but when the number of observation satellites is small,the speed position group The combining accuracy is lower than that of the pseudo range pseudo range rate method.Based on this,in order to better compare and study the effect of two coupled navigation systems on the same platform and reduce the calculation,this paper proposes to establish a combined switching mode,which uses the speed position combination scheme when the number of satellites is large and the pseudo range pseudo range rate combination scheme when the number of satellites is small.Fourth,due to the fact that the GNSS signal is easy to be interfered and blocked in the process of carrier movement,aiming at the problem that the accuracy of the combined system of GNSS/MARG decreases rapidly when the GNSS signal fails,the Elman neural network model is constructed.When the GNSS signal is good,the MARG error model is learned,and when the signal is blocked,the Elman neural network is used to predict and compensate the error of the MARG,so as to make the GNSS signal more accurate.Compared with the traditional BP neural network,the accuracy of the coupled navigation is improved by 80%.Finally,in order to verify the effectiveness of the scheme,the actual data collection platform is built and the data collection experiment is completed.The data processing results show that compared with the traditional GNSS/INS coupled navigation scheme,the GNSS/MARG coupled navigation scheme designed in this paper has a 40% improvement in heading accuracy and a 10% improvement in position and speed accuracy,and can keep the position error less than 10 m when the GNSS signal is blocked for 500 s,with remarkable improvement effect.