Precise Positioning By Integrating Android Built-in GNSS/IMU/Magnetometer Sensors

The development of the Global Navigation Satellite System(GNSS)has brought many conveniences to our daily lives.Before the release of the Android 7 system,developers can only obtain the user's PVT(Position Velocity and Time)information through the relevant Android Application Programming Interface(API)officially released by Google.That means Google officially closed the raw GNSS observations to the developers and only opened some GNSS auxiliary information(altitude angle,azimuth angle,signal-to-noise ratio,PRN,etc.)and satellite status information.In2016,with the release of Android 7 system,Google officially announced that developers can use raw GNSS observations from Android smart devices through APIs at the application level.That means that researchers can conduct research related to high-precision positioning and navigation on Android GNSS.At the same time,it also makes it possible to do research on Android GNSS and Android smart terminal multi-sensor data fusion.In this thesis,based on the above background,quality assessment of the Android GNSS raw observations is carried out.Then based on the Android GNSS raw observations,an improved Hatch Filter algorithm towards sub-meter positioning using only Android raw GNSS measurements without external augmentation corrections is discussed.Meanwhile,a data fusion positioning algorithm based on multi-sensor data and Android GNSS from Android smart terminals is discussed and analyzed.Both of the two algorithms are verified by relevant experiments to test their positioning performance and feasibility of the algorithms.The main works and contributions of this thesis are as follows:(1)Introduced the current market demand for Android smart terminal on location based on service applications and the relevant background of the availability of Android GNSS raw observations.Meanwhile,how to obtain the raw Android observations is introduced in detail.(2)Quality assessment of the Android GNSS raw observations is carried out from the aspects of signal-to-noise ratio,cycle slip and duty cycle,pseudorange and carrier phase noise levels etc.Restricted by their linear polarized antennas,the signal-to-noise ratio measurements of the mainstream Android smart terminals are about 10 d B-Hz lower than that of geodetic receivers.In addition,the Android smart terminal GNSS antenna has weaker suppression of multipath effects,and the antenna gain is uneven,resulting in its fast signal-to-noise ratio change with time.The rate of cycle slip of the mainstream Android smart terminal is higher than that of the geodetic receivers.Especially when the duty cycle is applied,the rate of cycle slip will have a higher magnitude.The noise level of the pseudorange of Android smart terminals is ten times than that of the geodetic receivers,while the noise level of the carrier phase is the same as that of the geodetic receivers.When the duty cycle is applied,the carrier phase measurements will be discontinuous caused by the duty cycle,which will cause the zero-baseline double-difference observations to be discontinuous.Therefore,the discontinuous carrier phase observations no longer have integer characteristics,so the ambiguity resolution cannot be fixed.(3)Carrier phase relative positioning performance analysis using Android raw GNSS observation.For the Android smart terminals affected by duty cycle,due to their discontinuous carrier phase observations,stable positioning accuracy cannot be guaranteed;for Android smart terminals without the introduction of the duty cycle,the convergence time and positioning accuracy are poorer compared to the geodetic receivers.However,in the open-sky environment,the positioning accuracy of carrier phase relative positioning by Android raw GNSS observations can achieve decimeter level under static and dynamic conditions.(4)TT-SD Hatch filter algorithm towards sub-meter positioning using only Android raw GNSS measurements without external augmentation corrections is proposed.Subsequently,the specific suppression effects of the Three-Thresholds and Single-Difference(TT-SD)Hatch filter algorithm on the ionospheric delay cumulative errors,cycle slips,and outliers,are verified by experiments.The feasibility of the TT-SD Hatch filter algorithm was verified by static and dynamic experiments,respectively.Under static and open-sky conditions,the horizontal and vertical position errors of TT-SD Hatch filter solution are about 0.6 and 0.8 m in terms of RMS,respectively;while under dynamic and open-sky conditions,the 2D error is about 0.9 m.That shows that its positioning accuracy is increasing by about 70% and64% comparing with the chipset solution.(5)Combining the position results calculated from the Android raw GNSS observations and the data of the built-in MEMS sensors(accelerometer,gyroscope,magnetometer)of Android smart terminals,Android GNSS/MEMS fusion positioning algorithm is proposed and verified.Allen variance is used analyze the error characteristics of the built-in MEMS sensors.Than the Android GNSS/MEMS fusion positioning algorithm is introduced in detail.Under dynamic conditions,the position results obtained by the fusion algorithm,which are more stable and robust,are consistent with the position result obtained by the position result only with GNSS methods.Experiments show that the 2D error of the dynamic RTK-INS and TT-SD Hatch Filter-INS methods are about 0.5m and 3.7m in a complex environment.Meanwhile,the accuracy of the fusion algorithms are increasing by 17.8% and 16.2%,respectively.