GNSS DATA PROCESSING. Volume II: Laboratory Exercises

Transcription

1 GNSS DATA PROCESSING Volume II: Laboratory Exercises

2

3 TM-23/2 May 2013 GNSS DATA PROCESSING Volume II: Laboratory Exercises J. Sanz Subirana, J.M. Juan Zornoza and M. Hernández-Pajares

4 Acknowledgements We would like to thank Dr Javier Ventura Traveset and Professor Dr Günter Hein for suggesting that we write this book and for their enthusiastic support and advice throughout the whole process. We also thank Carlos López de Echazarreta Martín for his help during the editing. We are grateful to Professor Dr Bernhard Hofmann-Wellenhof for his advice and encouragement. Special thanks go to Jordi Català Domínguez for redrawing most of the figures in this book. Our thanks also go to Adrià Rovira García for his continuous help and the deep implications of his review of this material, to Dagoberto Salazar for initial reviews of the English and to Georgina Sanz for her help and support with the editing of this book. We are also grateful to ESA Communications for their supporting advice and cooperation. Any comments regarding this book may be addressed to Cover: Galileo is the European GNSS (ESA/J. Huart). To Georgina Sanz Pons This book is for you An ESA Communications Production Publication GNSS Data Processing, Vol. II: Laboratory Exercises (ESA TM-23/2, May 2013) Production Editor Editing Publisher ISBN K. Fletcher Contactivity bv, Leiden, the Netherlands ESA Communications ESTEC, PO Box 299, 2200 AG Noordwijk, the Netherlands Tel: Fax: (two volumes plus CD) ISSN Copyright 2013 European Space Agency

5 Preface Preface This book is inspired by the welcome package that we offer to our doctoral students when they start their activities in the research group. The original package has been updated and compiled into two volumes which contain a self-learning course and software tools aimed at providing the necessary background to start work in an operative way in Global Navigation Satellite System (GNSS). The design and contents are focused on the instrumental use of the concepts and techniques involved in GNSS navigation and it is intended to include all the elements needed to understand how the system works and how to work with it. In this way, after working through the two volumes, the students should be able to develop their own tools for highaccuracy navigation, implementing the algorithms and expanding the skills learned. The first volume is devoted to the theory, providing a summary of the GNSSs (GPS, Glonass, Galileo and Beidou) fundamentals and algorithms. The second volume is devoted to laboratory exercises, with a wide range of selected practical examples going further into the theoretical concepts and their practical implementation. The exercises have been developed with a specialised software package provided on a CD-ROM together with a set of selected data files for the laboratory sessions. This is an end-to-end GNSS course addressed to all those professionals and students who wish to undertake a deeper study of satellite navigation, targeting the GNSS data processing and analysis issues. Starting from a review of the GNSS architecture, the contents of Volume I range from the analysis of basic observables (code pseudorange and carrier phase measurements) to setting up and solving the navigation equations for Standard Point Positioning (SPP) and Precise Point Positioning (PPP). It involves, in particular, an accurate modelling of GNSS measurements (up to the centimetre level of accuracy or better) as well as the required mathematical background to achieve the high-accuracy positioning goal. Parameter estimation techniques such as least squares and Kalman filtering are explained from a conceptual point of view, looking towards implementation at an algorithmic level. iii

6 TM-23/2 For this self-contained educational package, we have tried not only to explain the theoretical concepts and provide the software tools, but also to illustrate the results and give the methodology on GNSS data processing. This is achieved in Volume II through guided examples developed in laboratory sessions using actual GNSS data files in standard formats. Moreover, some additional practical information on public domain servers with GNSS data products (precise orbits and clocks, ionospheric corrections, etc.) is included, among other items. Most of the algorithms introduced in the theory are implemented in the GNSS-LABoratory (glab) tool suite. glab is an interactive and user-friendly educational multipurpose software package for processing and analysing GNSS data to centimetre-level positioning accuracy. The use and functionalities of this tool are thoroughly explained and illustrated through the different guided exercises in the laboratory sessions, together with self-explanatory templates, tool tips and warning messages included in the Graphical User Interface (GUI). glab has been developed under an European Space Agency (ESA) Education Office contract. The glab tool suite is complemented with an additional software package of simple routines implementing different algorithms described in the theory. 1 These elementary routines are included as examples of basic implementations to help students build their own tools. As mentioned above, the target is to provide effectiveness in instrumental use of the concepts and techniques of GNSS data processing from scratch. The didactic outline of this book is the result of more than 25 years of university teaching experience. In a similar way, the scientific/technological approach has been enhanced by our experience in developing different R&D projects in the GNSS field. 1 The glab source code is also provided as an example of the full implementation of the algorithms in a self-contained tool with a wide range of capabilities. iv

7 Contents Contents How to use this book vii 1 The glab Tool Suite 1 Session 1.1 Examples of Standard and Precise Positioning with glab 5 2 Laboratory Environment and Data Files 31 Session 2.1 UNIX Environment, Tools and Skills Session 2.2 GNSS Standard File Format Satellite Orbits and Clocks 53 Session 3.1 Coordinate Frames and Satellite Orbits Session 3.2 Errors in Orbits and Clocks of GPS Satellites Session 3.3 Glonass Broadcast Orbit Integration Measurement Analysis 113 Session 4.1 Code and Carrier Measurement Analysis Session 4.2 Code Carrier Measurements and Combinations of Three-Frequency Signals Session 4.3 Combinations of Trios of Carrier and Code Measurements Measurement Modelling and Positioning 167 Session 5.1 Standard and Precise Point Positioning with glab Session 5.2 Model Component Analysis for SPP and PPP and Implementation of Algorithms Session 5.3 Model Component Accuracy Assessment for GPS SPP 225 v

8 TM-23/2 6 Analysis of GPS SVN49 Anomaly 245 Session 6.1 Analysis of GPS SVN49 Anomaly A GNSS Elemental Routines and glab Libraries 263 Session A.1 Examples of GNSS Elemental Routines Session A.2 Programming with glab Modules B CD-ROM Content and Software Installation 307 B.1 CD-ROM Content B.2 Software Installation C Software Programs Associated with this Book 309 D Sources of Data Files Used in the Exercises 315 List of Acronyms 319 Bibliography 323 Index 325 vi

9 How to use this book How to use this book This volume contains the practical part of the GNSS course developed in these books. The fundamentals and algorithms appear in Volume I. The aim is to provide an experimental understanding of the theoretical basis and algorithms already introduced in the fundamentals and, along with this experimentation, to develop a methodology on GNSS data processing and analysis. This is achieved with a set of 15 practical sessions, distributed in six chapters, plus two appendices, and a specialised software package provided in a CD-ROM together with selected GNSS data files in standard format. The first chapter introduces the GNSS-LABoratory (glab) tool suite, which is an interactive and user-friendly software tool allowing high-accuracy positioning. The use and functionalities of this tool are thoroughly explained and illustrated in different guided exercises. These exercises are motivated by selected examples illustrating GPS performances in different scenarios, including different geometries of satellites in view, high ionospheric activity conditions or satellite clock anomalies, among others. The second chapter is devoted to describing the laboratory exercises environment and to providing an introduction to the GNSS standard data file formats. The UNIX (Linux) Operating System (OS) has been chosen as the baseline platform for the laboratory sessions, due to its high-performance and professional environment for dealing with these topics. For newcomers to this OS, a brief session on UNIX is included to introduce the basic commands and two additional software resources, awk/gawk and graph.py, widely used in the exercises. 2 Finally, the different GNSS standard file formats are explained in a user-friendly way through a set of HTML files with self-explanatory tool tips. The third chapter focuses on GNSS satellite coordinates, clocks and reference frames. Satellite coordinates and clock synchronisation errors are analysed using both broadcast navigation message RINEX files and precise orbit and clock SP3 files. GPS satellite coordinates are computed using the broadcast pseudo-keplerian orbital elements. Glonass orbits are integrated from the broadcast initial conditions, using a fourth-order Runge Kutta method. Several related issues are studied, such as short- and long-term accuracy, integration step width and different perturbations on the satellite orbits. The variation of osculating elements due to Earth s oblateness or Sun Moon acceleration is analysed from an experimental point of view. The accuracy of 2 Note: glab, awk and graph.py are also available for the Windows OS. vii

10 TM-23/2 the broadcast orbits and clocks is assessed against precise products, taking into account the different aspects involved, such as the reference frames and antenna phase centres used, coordinate interpolation, etc. Singular events, such as the GPS selective availability switch-off on 2 May 2000, or the Glonass reference frame transition from PZ-90 to PZ on 20 September 2007, are also studied with actual data from files collected during such periods. The fourth chapter targets GNSS measurements. Using actual measurements from RINEX files gathered from public domain data servers, the pseudorange (code) and carrier phase observables, as well as their different combinations (geometry-free, ionosphere-free, wide-lane, GRAPHIC), are analysed, and some of their characteristics directly from graphical results shown. Files collected under different ionospheric conditions (low solar activity, solar maximum, etc.), including special events such as the Halloween storm in October 2003, are processed to illustrate system performances in different scenarios, including extreme phenomena. The propagation of travelling ionospheric disturbances is also depicted from actual GPS measurements, in a very simple and straightforward way. Other exercises are devoted to the analysis of multipath and receiver noise in measurements and in combinations of measurements of two- and three-frequency signals. This study is carried out with several GNSS (GPS, Glonass and Galileo) data sources, including GPS measurements collected with anti-spoofing conditions both enabled and disabled. Different combinations of three-frequency signal measurements are analysed, assessing the noise and discussing their suitability for navigating or estimating ionospheric delays, as well. The feasibility of depicting the secondorder ionospheric effect is also explored and discussed using combinations of carriers from three-frequency signals. The fifth chapter deals with measurement modelling and positioning. The different terms involved in the code and carrier measurement modelling for SPP and PPP are studied in detail. Geometric ranges, relativistic correction, atmospheric effects (ionosphere and troposphere), instrumental delays, windup, solid tides, etc., are analysed and their magnitude and impact assessed on range and user domains. Using precise IGS products as reference values and GNSS measurements, the error budget in the SPP model is assessed and the transfer of range domain errors to the position domain is analysed. As mentioned above, the use and functionalities of glab are explained using the exercises in the laboratory sessions. As an example, a deep study of the computation of the static PPP solution for an IGS permanent receiver site is undertaken. This covers checks on all the products used (SP3 orbits and clocks, and ANTEX files), models and parameter settings (in the Kalman filter) applied to achieve a centimetre level of agreement between the IGS solution for the receiver coordinates and clock (SINEX file), as well as zenith tropospheric delay estimates. Other examples illustrating glab capabilities to process trajectories are illustrated with the kinematic positioning of a low Earth orbit satellite and an aircraft flight. viii

11 How to use this book To explain the implementation of the algorithms in detail, the GPS measurements are modelled by hand from scratch for the SPS, using simple routines for elemental functions (Klobuchar, troposphere, etc.), when needed. Afterwards, the navigation equations system is proposed and solved with Octave or MATLAB, using least squares and Kalman filtering. The laboratory sessions end with a work of synthesis presented in the sixth chapter. It is inspired by the paper by [Springer and Dilssner, 2009] published in the Inside GNSS journal and motivated by the analysis of the GPS SVN49 anomaly on the L1 signal. Two additional laboratory sessions are given in Appendix A, as complementary background, to help the reader develop his or her own GNSS software tools. A set of examples of elemental routines implementing basic functions (Klobuchar model, tropospheric model, satellite coordinates computation, among others) for end-to-end single-frequency GPS positioning are analysed in detail in the first session. Examples of dual-frequency-based cycle-slip detectors, among single-frequency detectors, are also given. The second session illustrates the use of glab as a library for developing GNSS software. Although a basic knowledge of the Linux OS is desirable, it is not essential to follow the book. By means of the different guided exercises, the reader is introduced in a natural way and by immersion to the syntax and possibilities of this environment. Our experience has shown that students with no previous knowledge of Linux, or the tools used in the exercises, do not experience any great difficulty in familiarising themselves with this environment; furthermore, they appreciate the fact that their training is applied to a real context where these problems are usually encountered. Nevertheless, and taking into account that the fundamental purpose of this publication is GNSS training, graphical results of the exercises have been included after each laboratory session, to develop many of the concepts of these sessions without the need to execute the programs. ix