Research On Android Smartphones’ Multi-GNSS Observation Data Quality And Positioning Algorithm

With the development of society,the demand for location service information in people’s daily life is constantly increasing,and smartphones that can provide location information will undoubtedly be relied on by most people.At present,the largest number of smartphone users is Android smartphones’.Global users have exceeded 2billion,and this number is increasing year by year.Android smartphones have rich positioning sensors,but the one most widely used is the GNSS positioning module.The GNSS positioning accuracy of Android smartphones is in the range of 4-8m,which can only meet the accuracy requirements of general users,and developers can only obtain information such as the position,speed from the PVT module.However,in 2016 Google company announced that Android 7 and later versions support GNSS raw measurements(including pseudorange,carrier,and Doppler measurements)output,this makes it possible to use Android smartphones to obtain a higher-precision position.Since then,many researchers at home and abroad have carried out many studies on the quality of GNSS measurements and the GNSS positioning algorithm based on Android smartphones.But most studies of Android smartphones positioning performance were carried out under good observation conditions,further research is need in more complex environments.This paper focuses on the research of positioning algorithm and GNSS data quality of Android smartphones positioning in different environments,the main work and conclusions of this paper are as follows:(1)This paper summarizes the research status of GNSS positioning of smartphones at home and abroad,and discusses the difficulties of using Android smartphones to obtain a continuous,reliable and high-precision position.the architecture of the Android smartphone GNSS module and the interface for acquiring GNSS data are studied;At the same time,the unified method and principle of the GPS/GLONASS/BDS/Galileo/QZSS time frame are explained;then the principles and methods of obtaining and generating pseudoranges,carrier phase,doppler observations through the GNSS data interface are explained.(2)This paper analyzed the GNSS data quality of Android smartphones in different environments from the aspects of satellite tracking number,dual-frequency data availability rate,data integrity rate,signal to noise ratio,multipath,cycle slip,observation noise and so on.The results show that: The number of satellites with dualfrequency observation data are far fewer than satellites with single-frequency observation data.The availability of dual-frequency data of GPS/QZSS、Galileo system is mainly limited by the number of satellites with the L5/E5 a observation data,but the pseudorange multipath in the L5/E5 a frequency band is much smaller than other system frequency bands;the signal-to-noise ratio of smart phones is mostly between 25-35 d BHz,and it has a weak correlation with the satellite altitude angle;GPS/QZSS L5 and Galileo E5 a band SNR is smaller than the corresponding L1/E1 band;The cycle slip of the carrier phase observation of the smartphone frequently occurs,and the cycle slip ratio of each frequency band is more than 5% even in an open environment;The noise of BDS B1/GPS L1/QZSS L1 pseudoranges is generally in the range of 1.5-1.6m;the noise of Galileo satellites is basically less than 1m;the pseudorange measurements noise of GLONSS system is between 3-5m;the noise of carrier phase observations of each system is less than 0.095 cycle.(3)Introdue the SPP/PPP single-frequency and dual-frequency positioning function model,parameter estimation method,random model and related error source correction models and methods.At the same time,the principle and method of singlefrequency/dual-frequency observations to detect cycle slip are introduced.Considering the large measurement noise of the pseudoranges,this paper studied the method of smoothing pseudoranges with Doppler observations and carrier phase observations,and on this basis,SPP positioning algorithm based on pseudoranges smoothed by the carrier phase / Doppler observations was implemented.At the same time,a set of PPP positioning algorithms suitable for smartphones was implemented.(4)In this paper,the SPP positioning results of the Android smartphone in different environments were researched,the main conclusions are as follows: The positioning accuracy of signal-to-noise ratio based stochastic model is better than that of satellite elevation angle based stochastic model.In the open environment,the carrierphase smoothed pseudorange can effectively improve the SPP positioning accuracy,but the effect is small under occluded areas;but the doppler smoothed pseudorange can effectively improve the SPP positioning accuracy in the open and occluded environments.For GPS/Galileo/QZSS satellites with dual-frequency observation data,the SPP positioning accuracy of GPS / QZSS L5 、 Galileo E5 a frequency band observation combination is the highest,followed by the GPS / QZSS L1、Galileo E1 frequency band,and the lowest is the GPS / QZSS L1 / L5、 Galileo E1 / E5 a dualfrequency combined ionospheric measurements.(5)Analysing the results of Android smartphones PPP positioning in different environments,we find that: In an open environment,the more satellites involved in the solution in the PPP solution,the convergence of PPP solution becomes quicker.But the single-system and dual-system combined positioning results appear unstable;Under severe occluded environments,even the positioning results of GPS/BDS/GLONASS/Galileo/QZSS combined system can not converge.In an open environment,the positioning accuracy of the GPS/BDS/GLONASS/Galileo/QZSS system combination is the highest,and the three-dimensional positioning accuracy can reach sub-meter level;the single-frequency PPP accuracy of the GPS/Galileo/QZSS combined system is higher than that of the dual-frequency ionospheric model PPP,which is mainly caused by the fact that there are few dual-frequency satellites,but as the observation time is extended,the dual-frequency PPP results appear to be very stable.Under severe occluded environments,the positioning accuracy of PPP has been improved exponentially compared to SPP,but still can’t meet the needs of highprecision applications.