Real-time GNSS/Accelerometer Fusion Algorithm And Its Application In Seismic Monitoring

GNSS real-time precise single point positioning technology can provide continuous positioning services in geological disaster monitoring,navigation positioning,and other fields.Compared with RTK technology,PPP technology has the advantage that it does not need density reference stations and is more convenient to maintain.However,due to the limitation of its long convergence time and relatively low precision,the real-time PPP technology is mainly used in a wide area of precision positioning services to provide continuous real-time monitoring,such as promoting the application of earthquake monitoring and early warning,However,when GNSS technology is applied in the field of seismic monitoring,it still has obvious shortcomings that is,the relatively low sampling rate leads to the weak ability of high-frequency signal acquisition and low precision.In order to make up for this deficiency,the monitoring performance can be improved by integrating GNSS and accelerometer data.Because the accelerometer has a higher sampling rate and measurement precision,it can quickly monitor and record the high-frequency strong vibration signal in the earthquake near-field.Therefore,the study of this paper focuses on improving the real-time fusion performance of GNSS and accelerometer.On the one hand,the author optimizes the real-time GNSS precise point processing technology.On the other hand,the author optimizes the fusion algorithm and realizes the real-time processing and the practical application of the instrument.The methods and technologies of real-time GNSS data processing are introduced in detail in this paper.In addition to the basic precise point positioning method,the performance of real-time GNSS precise point positioning can be improved by adopting multi-GNSS data processing and ambiguity resolution strategy,so as to accelerate convergence and improve positioning precision.The real-time GPS/GLONASS dual-system data processing is realized in this paper.The float solution without ambiguity resolution could achieve the convergence within 22.6 minutes averagely and the positioning accuracy in the east,north,up components are 2.2,1.2,4.2,respectively.After fixing GPS ambiguities in real-time,the results achieved the convergence within 15.3 minutes averagely,and the positioning accuracy in the east,north,up components reached 1.5,1.3,3.8,respectively.However,the method of GLONASS ambiguity resolution is for ionosphere-free ambiguity which has a shorter wavelength,so it is also more difficult to acquire good fixed results in real-time processing.Therefore,the results of fixing GLONASS ambiguity are not very stable.Therefore,this paper recommends that we can only fix the GPS ambiguity in real-time processing.In this paper,the integrated methods are introduced.The real-time GNSS and accelerometer tight-integrated and loose-integrated processings are realized respectively.The difference between the positioning results of the tight-integrated processing and loose-integrated processing is not big.Through the comparison of the vibration experimental results of tens of seconds,the difference between the results of the tight-integrated processing and the loose-integrated processing relative to the reference values is less than 1mm,but the processing efficiency of the loose-integrated processing is more than 10 times that of the tight-integrated processing.Therefore,loose-integrated processing is more suitable for real-time processing.In addition,real-time integration may cause discontinuities in the results due to the large sampling interval of GNSS and the noise of accelerometer.This deficiency can be improved by increasing the sampling rate of GNSS or smoothing the results using the backward filter.In this paper,near-real-time smoothing results are used for optimization,and a 5-second sliding window will lead to a 5-second delay.However,compared with the results of forward-filtering,the accuracy of standard deviation can be improved by more than 30%,which is of great help to the final fusion effect.The real-time loose-integrated processing algorithm is applied to the SIO geodetic module and MEMS accelerometer instrument package and the integrated GNSS seismograph instrument:SMAG2000 developed by Hubei Earthquake Agency and GNSS Research Center of Wuhan University.It can realize the online data synchronization acquisition,real-time processing on the server and online in-sta processing.The vibration experiments of the two instruments were compared with Trimble SG160-09.According to the results,SG160-09 was mainly affected by long-period fluctuations,which is mainly related to the GNSS processing effect.In our experimental site,SG160-09 had some deficiencies,but it used 10 Hz GNSS and 200 Hz acceleration data,its forward-filtering results were still very smooth.We compared different results in the time domain and the frequency domain.For the two instruments processed by our software,we output real-time forward-filtering results and near-real-time smoothing results,respectively.Among them,the errors of forward-filtering results in high frequency and low frequency were large,which were relatively worse than SG160-09,but this is because these two instruments used the 5Hz GNSS data.The near-real-time smoothing results improved the results significantly,which are the best of all other kinds of results.The standard deviation of the differences between the output results and the reference values could be less than 1 mm and the correlation coefficient of two series could be as high as 0.99.The results show that the algorithm of real-time GNSS processing and real-time fusion processing with accelerometers in this paper can be used in actual seismic monitoring successfully.