Vol. 1, No. 1, 2002

JournalJournal ofof GlobalGlobal PositioningPositioning SystemsSystems

ISSN 1446-3156 (Print Version) ISSN 1446-3164 (CD-ROM Version)

International Association of Chinese Professionals in Global Positioning Systems (CPGPS) Journal of Global Positioning Systems

Aims and Scope The Journal of Global Positioning Systems is a peer-reviewed international journal for the publication of new information, knowledge, scientific developments and applications of the global navigation satellite systems as well as other positioning, location and navigation technologies. The Journal will include original research papers, review articles and invited contributions. Short research and technical notes, book reviews, and commercial advertisements are also welcome. Specific questions about the suitability of prospective manuscripts may be directed to the Editor- in-Chief.

Editor-in-Chief Jinling Wang The University of New South Wales, Sydney, Australia [email protected]

Editorial Board Ruizhi Chen Changdon Kee Salah Sukkarieh Finnish Geodetic Institute, Finland Seoul National University, Korea The University of Sydney, Australia Wu Chen Hansjoerg Kutterer Todd Walter Hong Kong Polytechnic University, Hong Kong Deutsches Geodaetisches Forschungsinstitut, Stanford University, United States Dorota Grejner-Brzezinska Germany Lambert Wanninger Ohio State University, United States Jiancheng Li Ingenieurbuero Wanninger, Germany Ren Da Wuhan University, China Caijun Xu Bell Labs/Lucent Technologies, Inc., United States Esmond Mok Wuhan University, China C.D. de Jong Hong Kong Polytechnic University, Hong Kong Guochang Xu Delft University of Technology, The Netherlands J.F. Galera Monico GeoForschungsZentrum (GFZ) Potsdam, Hans-Jürgen Euler Departamento de Cartografia FCT/UNESP, Brazil Germany Leica Geo-systems, Switzerland Günther Retscher Ming Yang Yanming Feng Vienna University of Technology, Austria National Cheng Kung University, Taiwan Queensland University of Technology, Australia Gethin Roberts Kefei Zhang Yang Gao University of Nottingham, United Kingdom RMIT University, Australia University of Calgary, Canada Rock Santerre Guoqing Zhou Shaowei Han Laval University, Canada Old Dominion University, United States Thales Navigation, United States Bruno Scherzinger Applanix Corporation, Canada Editorial Advisory Board Junyong Chen Gerard Lachapelle Peter J.G. Teunissen National Bureau of Surveying and Mapping, China University of Calgary, Canada Delft University of Technology, The Netherlands Yongqi Chen Jingnan Liu Sien-Chong Wu Hong Kong Polytechnic University, Hong Kong Wuhan University, China Jet Propulsion Laboratory, NASA, United States Paul Cross Keith D. McDonald Yilin Zhao University College London, United Kingdom NavTech, United States Motorola, United States Guenter Hein Chris Rizos University FAF Munich, Germany The University of New South Wales, Australia

IT Support Team & Positioning Group The University of New South Wales, Australia

Publication and Copyright The Journal of Global Positioning Systems is an official publication of the International Association of Chinese Professionals in Global Positioning Systems (CPGPS). It is published twice a year, in June and December. The Journal will have both print version (ISSN 1446-3156) and CD-ROM version (ISSN 1446-3164), which can be accessed through the CPGPS website at http://www.cpgps.org/journal/journal.html. Whilst CPGPS owns all the copyright of all text material published in the Journal, the authors are responsible for the views and statement expressed in their papers and articles. Neither the authors, the editors nor CPGPS can accept any legal responsibility for the contents published in the journal.

Subscriptions and Advertising Membership dues in the CPGPS include the subscription to the Journal CPGPS Logo Design: during the period of membership. Subscriptions from non-members Peng Fang University of California, San Diego, United States and advertising inquiries should be directed to:

CPGPS Headquarters Cover Design and Layout: Department of Geomatics Engineering Satellite Navigation & Positioning Group The University of Calgary The University of New South Wales, Australia Calgary, Alberta, Canada T2N 1N4 Fax: +1(403) 284-1980 Printing: E-mail: [email protected] Wuhan University, China © CPGPS, 2002. All the rights reserved. Journal of Global Positioning Systems

Vol. 1, No.1, 2002

Table of Contents

The Modernization of GPS: Plans, New Capabilities and the Future Relationship to Galileo Keith D. McDonald ...... 1

Precise Ionosphere Modeling Using Regional GPS Network Data Y. Gao and Z.Z. Liu ...... 18

Multipath Mitigation for Bridge Deformation Monitoring G. W. Roberts, X. Meng, A. H. Dodson, E. Cosser...... 25

3D Multi-static SAR System for Terrain Imaging Based on Indirect GPS Signals Yonghong Li, Chris Rizos, Eugene Donskoi, John Homer, Bijan Mojarrabi...... 34

Accuracy Performance of Virtual Reference Station (VRS) Networks Günther Retscher ...... 40

Pseudolite Applications in Positioning and Navigation: Progress and Problems J. Wang...... 48

Experts Forum

An Overview of Atmospheric Radio Occultation T. P. Yunck,...... 58

The Contribution of GPS Flight Receivers to Global Gravity Field Recovery Peter Schwintzer and Christoph Reigber, ...... 61

Ocean Remote Sensing with GPS Cinzia Zuffada ...... 64

GPS For Ionospheric Sensing: Space and Ground Based E. A. Essex ...... 66

Journal of Global Positioning Systems (2002) Vol. 1, No. 1: 1-17

The Modernization of GPS: Plans, New Capabilities and the Future Relationship to Galileo

Keith D. McDonald Navtech Consulting, Alexandria, VA USA

Received: 18 July 2002 / Accepted: 18 July 2002

Abstract. This paper reviews the development, status that could be implemented with new generations of GPS and current capabilities of GPS. The modernization replenishment and follow-on spacecraft. This paper improvements planned for GPS are then discussed addresses the concerns, options, issues and plans during and summarized, including brief descriptions of the the next fifteen years and beyond for improvements to additional features planned for the spacecraft, the GPS and the significant performance benefits that will be available to users. The European Community is planning control segment and the user equipment. A to deploy a navigation satellite system with similar discussion is presented of the impact of the system performance characteristics to GPS in the 2008 time modernization plans and activities in improving the frame. We will briefly investigate the benefits and performance of the four principal operating modes features of the combined capabilities of GPS and Galileo of GPS. The implications of the GPS modernization that may impact international users in the future. and enhancement activities and their relationship to GPS has become an essential part of the navigation, the analogous European Galileo program activities positioning, surveillance and timing aspects of ground, and other GNSS efforts are covered. Several marine, aviation and space applications. The current uses, technical, policy and implementation concerns with new ones, will continue to grow resulting in a need relating to the timely deployment of the for even more demanding capabilities. improvements to GPS are briefly addressed.

2 Background Key words: GPS, Modernization, GNSS, GALILEO

2.1 Development and Implementation

The US Department of Defense (DOD) developed the 1 Introduction concept and general configuration for GPS in the early 1970's as a joint program involving all three military During the past three decades, the Global Positioning departments. The program was initially directed by the System (GPS) has grown from a navigation concept Joint Services Navigation Satellite Executive Steering through development and implementation to an Group (NAVSEG) formed in the Pentagon in 1968. This operational system of 28 spacecraft currently serving group was chartered to determine the feasibility and millions of users. Its use has increased such that over a practicality of a space-based navigation system for million GPS receivers a year have been produced since improving military navigation and positioning. The group 1997. The rapidly growing GPS market, including met regularly for over three years. It was to prepare, if equipment and applications, has been reliably estimated appropriate, a Development Concept Paper (DCP) (USGAO, 1998) at about $8.5B in 2000, to about $17B describing the technology, the performance capabilities, by 2003 and is expected to be in excess of $60B in 2010. the principal development areas, the cost, the benefits and the overall funding requirements for the system. The DCP The GPS has performed extremely well but a number of would then be presented to the Defense Systems desired and needed improvements have been identified

2 Journal of Global Positioning Systems

Acquisition Review Committee for consideration as a with the first operational (Block II) spacecraft launched DoD development program. on February 22, 1989. It was recognized early that the use of satellite systems, The basic signal characteristics for GPS that we have solid state electronics, evolving digital computers and today and for which millions of receivers have been related technologies could possibly provide significant designed and produced, were basically established in the enhancements to the performance capabilities of early to mid-1970's. Fortunately, the GPS system concept navigation users. The NAVSEG was supported by the parameters developed by the DoD Navsat Steering Group DoD Navigation Satellite Management Office and the and investigated in detail, as well as developed further laboratories of the military departments. The author was and implemented, by the GPS JPO and their contractors, Scientific Director of the DoD Navsat Program during have performed very well. However, it is clear that nearly this period that included serving as Executive Director of all users, both civil and military, can now benefit the NAVSEG and Chairman of the Navsat Management substantially from various GPS system enhancements, Office. modifications and additions. In 1973, after over three years of intensive technology After the tremendous technology advances during the last investigations, concept development, requirements quarter century, it appears reasonable to assess the current analyses and program definition efforts by the Steering and future capabilities of GPS in the context of these Group and others, the system was approved by the developments. This has become apparent to many. Both DSARC (representing the military departments) and the civil and military Committees and other activities in the Director of Defense Research and Engineering for past several years have strongly recommended advanced development and test. modernizing GPS. Support for this effort has also come from the highest levels of government (Gore, 1999). With the approval of the GPS program, the US Air Force was designated the Executive Agent for managing the implementation, and the GPS Joint Program Office (JPO) was established at the USAF Space and Missile Systems 2.2 Current Status Organization in Los Angeles. Ten GPS Block I developmental spacecraft were built under contract by Figure 1, entitled GPS Today illustrates the main Rockwell Space Division (now Boeing) and successfully elements of GPS as well as the frequencies, signals and launched from Vandenberg AFB between early 1978 and signal spectrum currently used by GPS. Table 1, entitled 1985. Tests with these spacecraft demonstrated the GPS Operational System Parameters and capabilities of the system and resulted in the approval by Characteristics summarizes the principal current GPS DOD for the implementation of an operational system, system characteristics.

• 24+ satellite constellation • ~ Half-geosynchronous orbits (10,900 nmi)

Time, orbit

Ground Monitor position and health Antennas Stations

(5)

(4) C/A Codes

+ Ascension P(Y)-Codes

FIX FOM 1 N 42* 01” 46.12” W 091* 38’ 54.36” Diego Garcia EL + 00862 ft

1 ON 2 3

me nu Kwajalein 456 7 8 9

WP T POS NAV

CLR 0 NUM MARK OFF LOC K

ZER O I ZE Hawaii Rockwell Master Control * Vandenberg Receiver Station L2 L1 calculates (Shriever AFB) 1227.6 MHz 1575.42 MHz 3-D location, Colorado Springs, CO ± 12 MHz ± 12 MHz velocity & time

Fig. 1 GPS Today

McDonald: The Modernization of GPS 3

The general characteristics of the ground segment are deployment is given in Figure 3 entitled GPS Space given in Figure 2 entitled GPS Ground Control System Segment.

and similar information for the GPS spacecraft (S/C)

GPS Constellaton

COLORADO SPRINGS

VANDENBERG

CAPE CANAVERAL

HAWAII

KWAJALEIN

Ground Monitor ASCENSION DIEGO

Stations GARCIA

Antennas

Master Control

Station (Schriever AFB)

„ Master Control Station (MCS): Satellite control, system operations „ Alternate Master Control Station: Training, back-up z Monitor Station (MS): L-band; Collect range data, monitor nav signals ▲ Ground Antenna (GA): S-band; Transmit data/commands, collect telemetry

Fig. 2 GPS Operational Control System (OCS)

Tab. 1 GPS Operational System Parameters and Characteristics

2.22.99kdm SpaceSpace Segment Control Segment Segment User Segment Segment

2424 GPSGPS spacecraftspacecraft (S/C)(S/C) inin orbitorbit Headquarters & Master Control Station Passive receivers; no user transmissions isis thethe baselinebaseline constellationconstellation (MCS)(MCS) atat SchrieverSchriever AFB,AFB, ColCol Spr,Spr, COCO required for navigation reception 66 orbitorbit planesplanes spacedspaced 6060°°apartapart (LAN)(LAN) Uplink and monitor station locations: Receivers normally process signals from 44 S/CS/C perper orbitorbit plane;plane; >4>4 forfor >> 2424 constelconstel Schriever AFB multiple (5-12) S/C simultaneously 10,89810,898 n.mi.n.mi. (20,180(20,180 km)km) orbitorbit altitudealtitude Hawaii (monitor only) Number of S/C in view of unobstructed 55°55° S/C S/C orbitorbit planeplane inclinationinclination Ascension Island GPS receiver: 6-12 S/C (average 8) Near circular orbits (e ≈≈ 0)0) Kwajalein 1111 hrhr 5858 minmin S/CS/C orbitalorbital periodperiod Diego Garcia Coarse acquisition (C/A)-codes, 1.023 Mbps Kennedy Ctr., FL (up-link-- backup) used normally by civil community and ProvidesProvides PRN-codedPRN-coded rangingranging signalssignals New monitor sites: by military for acquisition of P/Y-code Circularly polarized transmissions 8 NIMA tracking stations FrequenciesFrequencies andand PRNPRN codes:codes: New M-code to provide direct access 1575.421575.42 ±± 1212 MHz:MHz: L1L1 bandband MCS receives monitor station measure-- 1227.61227.6 ±± 1212 MHz:MHz: L2L2 bandband ments, 2f iono corrections, UTC time Spread spectrum signal; pseudorandom- 1176.451176.45 ±± 1212 MHz:MHz: L5L5 bandband Large Kalman estimator at MCS noise coded: 1.023 Mbps & 10.23 Mbps 1.0231.023 MbpsMbps == C/A-codeC/A-code bitbit raterate MCS computes and schedules uplink 37 PRN C/A Gold codes available 10.2310.23 MbpsMbps == P/Y-codeP/Y-code bitbit raterate transmissionstransmissions thatthat include:include: Direct sequence P/Y-code: 37+ weeks long; 11 msec msec = = periodperiod ofof C/A-codesC/A-codes S/C ephemeris (orbital elements) 1 week segment used for each S/C 77 daysdays == periodperiod ofof P/Y-codesP/Y-codes (each(each S/C)S/C) S/C atomic clock corrections S/C almanac data GPS receiver antennas normally are for 11 GPSGPS S/CS/C perper launchlaunch (Delta(Delta II)II) Ionospheric delay model terms L-band upper hemisphere coverage C/A-codes in the clear (SA removed) S/C health and status information Signal power from S/C at receiver: P/YP/Y andand M-codes,M-codes, securesecure (encrypted)(encrypted) -160 dBw for L1 C/A-code signal Data message: 50 bps (mod 2 encoded) Monitor stations linked to MCS -163dBw for L1 P/Y- code signal S/CS/C lifetimes:lifetimes: II,II, IIAIIA-7.5- 7.5 years;years; Up-link antennas: 10 m. dishes -166dBw for L2 P/Y- code signal IIRIIR -7.5 -7.5 +years;+years; IIF:IIF: 12-1512-15 yearsyears Near continuous ant. visibility of GPS S/C Modernized L1, L2 and L5 signals are Unified S-band-- tracking, telemetry & planned for current L1 C/A pwror more S/C:S/C: MultipleMultiple RbRb and and CsCs atomicatomic clocksclocks command (TT&C) links Civil GPS standalone accuracy:-- 5 - 10 m. Differential GPS accuracy: 1 mm - 3 m. Military GPS accuracy:-- 2 – 9 m. Orbit and ephemeris accuracy (OCS): Digital correlation receivers are small, 1414 -- 180 180 dayday S/CS/C nav.nav. messagemessage storagestorage lightlight andand reliable.reliable. 8M+8M+ inin serviceservice (1/02)(1/02)

4 Journal of Global Positioning Systems

Block II/IIA

Block IIR

Block IIF

• 24-satellite (nominal) constellation

• Six orbital planes, four satellites per plane

• Semi-synchronous, circular orbits (~10,900 n. mi., 20,200 km.)

Fig. 3 GPS Space Segment The Block II operational spacecraft launches which began booster failed on its newly configured Delta II launch in early 1989 later incorporated slightly modified (Block vehicle. A later launch in July of 1998 and all subsequent IIA) spacecraft. These included some improvements launches have been successful. As of 1998, all 28 of the including additional on-board data memory, providing an initial Block II (and IIA) operational spacecraft have been extended period (several months in lieu of 14 days) over launched. The remaining Block IIR spacecraft, including which data could be transmitted from the S/C to users nine to twelve S/C planned for modernization, will phase without ground uploads. in to provide the principal operational signals and capabilities for the system during the next ten years or The DOD contracted with Rockwell Space Division for more. 28 Block II and IIA spacecraft at a cost of about $1.2B. In July 1995, the full operational capability of GPS was The DOD has contracted with the Boeing Space Division achieved, consisting of a ground segment, a constellation for a fourth generation follow-on GPS (Block IIF) of 24 operational GPS spacecraft providing navigation spacecraft. Although a buy of 30-33 spacecraft was services worldwide and a variety of user equipment. originally planned, modernization requirements have Since that time, the GPS space segment has performed resulted in the need for a new generation of spacecraft continuously with between 24 and 28 operational beyond the IIF. To this end, the GPS Block III’s are in spacecraft. User equipment development and development. The Block III’s will replace all but twelve manufacture has increased dramatically, especially in the of the originally planned Block IIF spacecraft. civil community. The first six IIF spacecraft are on contract and an additional six are planned for GPS mid-term constellation sustainment purposes. Beyond that, the new Block III 2.3 The GPS Constellation spacecraft will be deployed. The Block IIR, IIF and III spacecraft provide the principal space vehicles available The Block II and IIA spacecraft's limited lifetime in orbit, for the enhancement and modernization of GPS. nominally about 7.5 years, establishes the need and schedule for replenishment spacecraft. The DOD It is the periodic replenishment of the spacecraft contracted with Lockheed-Martin Astro-Space for constellation that provides the opportunity for twenty-one GPS third generation replenishment (Block implementing upgrades and modernization features to the IIR) spacecraft. system. Since the lead time for acquiring a new generation of GPS spacecraft is long (typically 5-8 years), In January 1997, the first of the Block IIR spacecraft was and the full deployment of a new constellation requires a launched but did not achieve orbit when a strap-on number of years (typically 7-10 years), the introduction

McDonald: The Modernization of GPS 5 of improvements to GPS is normally a slow and gradual user receivers. Since the ionosphere is a dispersive process. medium with close to a 1/f2 dependence, the combination of the L1 and L2 frequencies provides an excellent real time technique for determining the group delay effects on 2.4 The Navigation Signals the GPS signal paths caused by the refraction of the signal paths by the free electron content in the GPS spacecraft currently transmit navigation signals to ionosphere. the earth at two frequencies, designated L1 and L2 (refer Ionospheric errors, if uncorrected, can contribute the to Figure 1). L1 is the principal GPS carrier signal, at a largest single propagation error to GPS operation, causing frequency of 1575.42 MHz. This signal is modulated by ranging errors of up to 40 meters. The GPS L2 signal was two types of pseudo-random noise (PRN) codes, termed originally implemented primarily for correction of the the coarse/acquisition (C/A) codes, at a bit rate (or slowly changing ionospheric delay. For this reason, it is "chipping" rate) of 1.023 Mcps, and the precision/secure transmitted at a lower power level (at about one fourth, or (P/Y) codes, at a chipping rate of 10.23 Mcps. -6dB) relative to the L1 C/A-code signal. The P/Y-code The GPS L2 signal, transmitted by the spacecraft at signals on L1 and L2 are normally secure, so the full 1227.6 MHz, was established to provide a second capabilities of these signals are accessible only to DOD frequency for ionospheric group delay corrections to GPS or other authorized users.

Civil Signals

Aeronautical Radionavigation Radionavigation L1 C/A-codes Services (ARNS) Band Satellite Services RNSS Bands (for civil & Current 960-1215 MHz (RNSS)Band 1215-1240 1560-1610 MHz military use)

Frequency in MHz 1575.42 L5 L2 Modernized I5, Q5-codes C-codes C/A-codes (new) (new) Frequency 1164.45 1188.45 1215 1227.6 1575.42 1176.45 (No data message on L5 quadrature signal) L1 Military Signals C/A-codes Current L2

Frequency in MHz P/Y-codes P/Y-codes 1227.6 1575.42 C/A-codes Modernized M-codes L2 M-codes Frequency

Note 1: Military M-codes are in definition. 1215.6 1239.6 1563.42 1587.42 L5 codes (I & Q-codes) are specified P/Y-codes P/Y-codes Note 2: Modulation envelopes for only the code principal signal lobes are shown. 1227.6 (L2) 1575.42 (L1) Note 3: Phase quadrature signals are shown below the origin lines.

Fig. 4 GPS Signal Evolution and Spectrum Occupancy signals transmitted by GPS and the planned evolution to GPS can also be used in a differential mode in which additional codes and signal frequencies. known navigation data or ranging data received at a reference location and time are compared (or differenced) with similar GPS measured data at the same point and 2.5 GPS Performance time. The corrections from this process are then applied at the same time (or with minimum latency) to the GPS GPS has a number of different modes of operation, each measured data taken at a remote point. with its own set of performance capabilities. First, GPS For real-time operation, a data link between the reference can be used autonomously (on a stand-alone basis), i.e. receiver and the remote receiver is normally used to without any augmentations. In this case, the user communicate the corrections. This process has the great equipment receives and uses only the signals received advantage of canceling the fixed, or slowly varying, from the constellation of spacecraft to determine user (bias) errors in measurement that have the same effect at position, velocity, time (PVT) and related parameters. both locations. The differential correction technique,

6 Journal of Global Positioning Systems which many other navigation systems have used to greater precision, the differential measurements of the improve performance, performs well with GPS since GPS carrier phase. most errors appear as slowly varying biases. The various modes of operation for GPS user equipment These differential corrections can be applied directly in and the corresponding nominal current performance real time or they can be stored, normally synchronized to capabilities are summarized in Table 2. The change in a common time source (such as GPS time) and employed GPS performance characteristics for various GPS stand- later using post processing methods. The corrections are alone and differential modes of operation for 2001 and usually established at the reference location and applied 2011 are discussed later. Performance accuracy values in at the remote (rover) receiver. Corrections can be based position, velocity, time and angle measurement (attitude) either on the GPS differential code measurements, or for are provided.

Tab. 2 GPS Performance and Modes of Operation With a Summary of Nominal Current Capabilities

Stand-alone Receivers • Autonomous with unaided code Exception (typical): carrier (Doppler) aiding of code to reduce code noise • Pseudorange performance: C/A-code ~0.5 - 2.0 m.; with narrow correlator ~10 - 40 cm. P/Y-code ~10 - 50 cm. • Environmental effects: ionosphere and troposphere delays • Other errors: ephemeris, multipath, receiver noise, S/C clock, receiver and S/C unmodeled delay errors • Overall nominal accuracy: C/A-code receivers ~ 5 - 10 m. P/Y-code receivers ~ 2 - 9 m. Differential (DGPS) Code Receivers • Environmental and other systematic effects: greatly reduced; limited by spatial decorrelation • Differential corrections: code measurement based; some latency effects • Low values of receiver code noise advantageous for improving measurement precision • Overall nominal accuracy: C/A-code DGPS receivers ~0.7 - 3 m., with narrow correlator ~0.5 - 2.0 m. P/Y-code DGPS receivers ~0.5 - 2.0 m. Real Time Kinematic (RTK) Carrier Phase Measurement (CPM) Receivers (DGPS) • Low noise measurement observable; differential phase error ~ 0.5 - 2 mm. • Differencing techniques employed (1,2,3∆) to remove unmodeled delay errors • Overall nominal accuracy: ~2 mm - 20 cm.; depends on baseline separation, observation period, corrections … C/A (or P/Y) – codes typically used but not essential; measurement observable is relative carrier phase Interferometric Measurement (carrier phase) Attitude Determination Receivers • Same general characteristics as CPM and RTK receivers • Nominal accuracy: ~1 mRadian; normally employs C/A (or P/Y) –codes but relative carrier phase is observable

and Hawaii, for accurate coastal, river, harbor and harbor entrance navigation for marine users. 2.6 Local Area and Wide Area Enhancements • The Wide Area Augmentation System (WAAS) of the Federal Aviation Administration (FAA) planned for To improve the performance of GPS, a number of local initial operation in 2002 to provide code corrections, area and wide area differential GPS (DGPS) systems and ionospheric delay and integrity data. networks have been established or planned. These include: • The FAA’s Local Area Augmentation System (LAAS), for use in precision landing of aircraft. Planned for • Local post processing and real time (kinematic) implementation by about 2003, this provides code systems for the GPS surveying community. corrections, integrity and other data. • FM radio sub-carrier transmissions of DGPS • The European Geostationary Navigation Overlay corrections for vehicular applications in many North System (EGNOS), similar to WAAS and planned for American and European cities. initial implementation in Europe for aviation uses by • Satellite-based transmissions of DGPS corrections about 2002. worldwide (by RACAL, Fugro and others) for use by • The Mobile Transportation Satellite System (MTSAT), surveyors, geodesists and geographic information planned by Japan for operation in a large region of the system (GIS) users. Pacific, providing code corrections and other data, for • The U.S.Coast Guard Differential Network, for operation in about 2002. continental US coastal regions, Gulf of Mexico, Alaska

McDonald: The Modernization of GPS 7

3 The Modernization of GPS Concerns on L2 and a New L5 The FAA for some time has opposed the use of L2 for Although GPS has performed extremely well and has aviation safety applications. Their concern is that since generally exceeded expectations, some significant the International Telecommunications Union (ITU) has improvements are needed. A number of committees, authorized this band for use on a co-primary basis with representing both government and civil communities, radiolocation services (including high power radars), that have investigated the system's needs and deficiencies aircraft using the band may be subject to unacceptable over the past decade in order to determine what levels of interference. Since this may compromise capabilities and features should be incorporated into a aviation safety operations, the FAA considers the L2 future GPS to satisfy both military and civil users. band unacceptable. The FAA requested a GPS aviation The modernization of GPS is a difficult and complex frequency in the Aeronautical Radionavigation Services enterprise. It involves not only addressing civil and (ARNS) band, which is located directly below the GPS military needs and costs for performance improvements, L2 band. As we will discuss, this has occurred. but also issues with far-reaching implications in other The Presidential Decision Directive (PDD) on GPS of areas. These issues include spectrum needs and use, March 29, 1996 (Gore and Pena, 1996) stated that both security, civil and military performance, system integrity, the L1 and L2 frequency bands would be available for signal availability, institutional concerns on GPS civil use and that a third civil signal would also be financing and management, and the future operation of authorized. Although the L1 and L2 frequency bands can GPS as a national and international resource. satisfy most civil users, aviation users need a third civil Fortunately, many of the critical issues have been frequency to replace the L2 band and its limitations in identified and are resolved or appear to be near final safety-of-life applications. resolution. If all goes as planned, or as hoped, it becomes After an intensive search for new frequencies by the reasonably clear what can be expected in the next decade. Department of Transportation (DOT), the Department of However, this assumes an optimistic view of the Defense (DOD) and other agencies, Vice President Gore commitment of the U.S. government to provide the announced on January 25, 1999 that a region in the decisions, the institutional arrangements and the funding ARNS band had been agreed upon as the new (third) civil necessary to meet the generally agreed upon needs in a frequency. This frequency is referred to as L5 and is timely manner. centered at 1176.45 MHz. This selection appears to satisfy aviation safety uses. 3.1 Signals and Signal Separation Figure 5 illustrates the arrangement in the spectrum now planned for the modernized signals, showing the ARNS Initially, the military GPS planners wanted to separate band below the GPS L2 band. their GPS frequencies from those used for civil Prior to operational use of L5, coordination with other applications. This separation was intended to avoid signal systems in the band as well as approval by the interference and interactions and to provide maximum International Telecommunications Union (ITU) is flexibility for both user groups. required. The ARNS band is currently established by the An intensive search, however, failed to find any new ITU for ground-to-air services. For part of this band to be spectrum that would satisfy the military requirements. used by GPS requires an ITU satellite-to-earth Therefore, the military decided that their signals will transmission classification. This matter was favorably remain within the current 24 MHz bands authorized by considered at the ITU World Radio Conference held in the ITU for GPS in the Radionavigation Satellite Services April, 2000 in Istanbul. It was expected that the (RNSS) band at L1 (1563.42-1587.42 MHz) and L2 international aviation community would support this (1215.6-1239.6 MHz). change. Additional Signal Needs It has been agreed for several years that additional GPS 3.2 New and Retained GPS Signals signals are needed for civil applications. These signals are required for a) reducing the ionospheric errors by use of For backward compatibility (or “legacy”) purposes, the the two frequency correction technique, b) for increased existing C/A-codes on L1 and the P/Y-codes on L1 and signal robustness, especially in aviation safety operations, L2 are to be retained. Continuation of these codes is and c) for improved acquisition and accuracy. necessary until modernized GPS spacecraft transmitting the new GPS signals for both civil and military users are deployed. New GPS user equipment also needs to be produced to operate with the modernized signals.

8 Journal of Global Positioning Systems

Band Assignments:

L1

RNSS - Radionavigation Satellite

RNSS Band Services Band

ARNS - Aeronautical Radionavigation GLONASS Services Band GPS L1 MSS JTIDS - Joint Tactical Information L1 Distribution System (DoD) 1605 GLONASS - Glonass Navigation Satellite 1563.42 1587.42 1595

System (Russian Federation) 1559 1575.42 1610

MSS – Mobile Satellite Services Band

L5 L2

ARNS Band RNSS Band

GLONASS

GPS L5 JTIDS… GPS L2 L2

1240 1164.45 1188.45 1250 1176.45 1215 1227.6 1239.6 1260

Fig. 5 GPS and GLONASS Signal Placement Arrangement: L1, L2 and L5 Signal Codes rate and longer codes than the current C/A-codes or the new C-codes. The proposed codes on L5 consist of a The current plans are to continue providing the civil 10.23 Mcps code rate with a code length of 10,230 bits community with the C/A-codes on L1, and to transmit (in one millisecond). This is the same code period as the newly configured C-codes (or sometimes referred to as C/A-codes. Compared to the C/A-codes, the longer, high CS-codes) on L2 from subsequent spacecraft when rate code sequences provides improved ranging accuracy, feasible. Plans for the L5-codes are for a higher bit (chip) a lower code noise floor, acceptable acquisition times,

1166.22 MHz 1186.68 MHz 1176.45 MHz

• Safety-of-life applications (e.g., civil aviation) • In Aeronautical Radio Navigation Service (ARNS) band • Two signals: (a) In-phase signal (I5) with data message (b) Quadrature signal (Q5); no data message for improved tracking • Higher accuracy when used with civil codes on L1 or L2 – Similar accuracy as military signals today (P/Y-codes) – More robust than current C/A-codes on L1 – Greater resistance to interference than C/A-codes – Higher chipping rate improves multipath performance – Plans are to have about four times greater power (needed to operate in ARNS band that has higher noise and interference levels) – Improved data message (and shortened)

Fig. 6 L5 - New Civil Signal

McDonald: The Modernization of GPS 9 better isolation (cross-correlation properties) between The final arrangement for the M-code signals appear to codes, and substantially reduced multipath interference use a secure M-code with a bit rate of about five Mbps susceptibility. Figure 6 entitled L5 – New Civil Signal modulated on dual "split spectrum" carriers that are shows the main spectral characteristics of the L5 in-phase spaced about 10 MHz above and below the centers of the and quadrature codes referred to as the I5 and Q5-codes. GPS L1 and L2 bands [at the first nulls in the current P(Y)-code structure]. The GPS JPO and their contractors To take full advantage of L5, it is planned for one of the have investigated the military signal alternatives and are two quadrature signals to be transmitted without data responsible for the final selection. modulation. The “data free” signal provides advantages for accurate phase tracking and more precise carrier phase measurements, of special interest to the survey and scientific communities. Similarly, the new C-codes on L2 4 Security Issues are at the C/A-code chipping rate (1.023 Mcps) and are time multiplexed to provide a data free signal as well as a Selective Availability signal with data. A number of Committees have investigated the need and The current P/Y-coded signals on L1 and L2 perform effectiveness of SA. Almost all of these, both civil and very well. However, because of the extremely long military, have found that SA is not effective, is easily sequence length (~1013 bits), and the corresponding mitigated (for example, by the use of differential period of this code (7 days), acquisition of the P/Y-code techniques) and that it is costly to continue, especially to is very difficult unless some precise knowledge of the civil users. Several of the groups strongly recommended system timing is known. The P/Y-code acquisition that SA be removed immediately (NRC, 1995; NAPA, normally involves first, the acquisition of the short 1995). sequence (1 ms) C/A-codes on L1. The C/A-code The Presidential Decision Directive on GPS released on message contains system timing data (in the hand-over, or March 29, 1996 stated that the continuance of SA would HOW, word) that provides an authorized user with be reviewed annually by the President starting in 2000 information for acquiring the P/Y-code. and that SA would be discontinued no later than 2006. However, the military believes it essential in the future to The civil community has consistently recommended that acquire their secure signals without first accessing the SA be removed. However, users adapted quickly to the C/A-codes. The C/A-codes are available to civil users and use of SA mitigation techniques, principally by the use of others. The military requirement for direct acquisition of differential corrections. their secure signal appears to require the transmission of a GPS Accuracy Improvements and Performance set of new military codes, called the M-codes. Augmentations Code Options Considered On May 1, 2000, the White House announced that the Initially, the DoD indicated that it planned to place its Selective Availability degradation would be removed new signals in the center of the L1 and L2 bands, similar starting at midnight, May 1, and that there were no plans to their C/A and P/Y-codes. To minimize interactions or intentions to restore it in the future. The substantial with the main part of the DoD signals, it appeared improvement in stand-alone GPS accuracy to civil users advantageous for the civil signals to avoid the center of has become apparent. Even at the time of the solar the bands. The author proposed on several occasions maximum effects on the ionosphere (in 2000-2001), during 1997-98 the use of pairs of coded signals (offset performance of SPS receivers is now typically at the 5-10 from each band center) in the L1 and L2 bands meter level. The removal of SA combined with the (McDonald, 1998a,b). The use of civil “separated availability of a second (and third) civil signal frequency carriers” or “split spectrum” signals in the outer regions can improve GPS stand-alone civil horizontal accuracy to of the bands where the military signals would be at a low the 1-3 meter range (at a 95% confidence level). The level provided separation between the military and civil removal of SA also has a substantial impact on reducing signals. It also provided some significant advantages for DGPS data link capacity requirements because of the precise carrier phase measurements and more precise reduced need for frequent SA corrections. code processing. Military Signal Security These split spectrum signals were then found to have Current military users normally access the C/A-coded excellent performance properties, especially compared to signals but their security system corrects for the effects of the current military P/Y-code signals, and to be easily the SA degradation. The DoD encryption of the P-code to implemented (Spilker et al., 1998). During the past two form the Y-code not only limits access to the most years the DoD has sponsored a number of investigations accurate codes to military (and other authorized) users of split spectrum signals and has selected this signal as but also provides anti-spoof (A-S) protection. The current their new M-code signal for their use at both L1 and L2. plan is for the P/Y-coded signals to be retained until the

10 Journal of Global Positioning Systems new M-code signals are generally available. The phase- performance and are planned to be directly accessed. The over interval can be expected to last until 2015 and military M-code signal structure is in final development. possibly longer. Figure 7 entitled New Military Signals : M-Code shows the general characteristics of the military P(Y) and the The new M-code signals will also be secure. They differ planned M-code signals. from the P/Y-code signals in that they provide improved

1215 1227 1239 1563 1575 Frequency in MHz 1587

• Anti-jam through higher power • Robust and autonomous acquisition • Spectral isolation from civil signals • Improved security (exclusivity, authenticity, confidentiality) • Better performance • Flexibility • Compatibility with C/A-code and P(Y)- code receivers • Operation within existing L1 and L2 bands

Fig. 7 New Military Signals: M-code improve the civil accuracy for stand-alone GPS users to the 1-3 meter level, or better. 5 GPS Accuracy Improvements and Performance Augmentations Unfortunately, the current schedule for the deployment of enough Block IIF spacecraft (about 18 are required) to Ionospheric Errors allow confident access to the L5 signals is over a decade away. A reasonable estimate for the date at which full Ionospheric propagation group delay effects on GPS operational reception of L5 signals would be about 2015, signals cause most of the residual receiver error. These depending on the lifetime of the IIA, IIR and IIF delay effects vary considerably depending on random spacecraft. effects, the time of day, season of the year and the activity state of the 11-year period solar (sunspot) cycle. Receiver Noise Errors Figure 8 entitled Solar Cycle 23 Sunspot Number Other factors that influence GPS user equipment accuracy Prediction provides a graphic indication of the cyclic include the number of parallel receive channels available character of the solar emissions that affect the earth’s to a receiver, the receiver’s code noise performance, its ionosphere. susceptibility to multipath, and the errors associated with The ionospheric propagation delay error effects can be implementing, or mechanizing, the solutions. The number effectively removed by using the two frequency of GPS receiver channels corresponds to the maximum ionospheric correction technique that the military now number of spacecraft that can be simultaneously tracked. uses. For this reason, there has been interest for many Ideally, the number of channels is large enough to years in a second frequency for civil GPS users. As provide continuous all-in-view tracking of the visible indicated earlier, agreement has been reached within the GPS spacecraft, since this provides the best geometric government to incorporate a second civil signal (at L2) performance. and a third civil signal (L5, in the ARNS band at 1176.45 Although the receiver code noise for a conventional C/A- MHz) in future Block IIF spacecraft. The excellent code receiver (with carrier aiding) is equivalent to a capabilities of these signals for ionospheric correction can ranging error of about one meter, code noise errors for

McDonald: The Modernization of GPS 11

Fig. 8 Cycle 23 Sunspot Number Prediction (September 200) "narrow correlator" C/A-code receivers can be as low as GPS tracking network and the Operational Control 10-20 cm. Implementation errors are usually small or System. negligible. Receiver thermal noise performance at the 10 Autonav Operation cm. to 1 m. level is considered good, however GPS precise carrier phase measurements are typically The GPS constellation may be required to operate considerably better, nominally at the 0.5-2 mm. without the GPS ground segment for an extended period. equivalent noise level. By accurately ranging to other spacecraft using the GPS inter-satellite link the Block IIR, IIF and III spacecraft Control Segment Errors can operate in an autonomous navigation (autonav) mode. The GPS control segment determines the quality of the The autonav ranging data is cross-linked to other GPS spacecraft orbital elements and timing data. These spacecraft in the constellation to provide continuous on- are uploaded to the GPS spacecraft memory and then board information that is used to autonomously compute periodically transmitted to the users in the GPS data accurate new ephemeris data for each spacecraft. This message. This spacecraft position and other data directly data, incorporating the measured GPS spacecraft orbital affect user accuracy. Moreover, since it degrades with perturbations, can provide excellent system accuracy over time, the data is influenced by the update rate of the an extended period (several months). uploads to the GPS spacecraft. Recent improvements in Position Accuracy Estimates for 2002 and 2012 the control segment provide spacecraft ephemeris accuracy at the 1-2 meter level. Figure 9, entitled Position Accuracy Estimates for Civil and Military GPS Receivers for 2000 to 2010 illustrates The planned addition of six ground stations of the the anticipated system performance improvements in National Imagery and Mapping Administration (NIMA) GPS as reflected in GPS receivers operating in their to the GPS tracking network will substantially improve various modes. This chart addresses position accuracy the quality and timeliness of the GPS tracking only, however there are many other measurements that measurements and the computed parameters. More will be affected as well. A brief summary of the frequent uploads to the spacecraft are also planned. In the additional measurement parameters and their estimated 2000-2010 period, sub-meter ephemeris accuracy, which accuracy is given in Table 3, entitled GPS Performance will improve to the decimeter range, is expected for the in Various Modes of Operation for 2000 and 2010.

12 Journal of Global Positioning Systems

Position Accuracy = Horizontal position accuracy at 95% confidence level. 2000 Year 2005 06 07 08 09 2010

100 m Current Civil GPS Systems 60 m Standard Positioning (SPS) Svc. 7 m Stand-alone C/A-code receiver (Current 1) 10 m Current 1 3-5 m. 6 m Degrades with solar cycle max at ~2011 C/A-code (1.023 Mcps) on 2 m L1 (1575.42 MHz) 1 m C2 30 cm 50 cm DGPS receiver using C/A-code (C2) L2 (1227.6 MHz) carrier phase C3 Carrier aided; separation of ~20-200km 10 cm ionospheric delay correction 2 cm RTK, Real Time Kinematic (C3); carrier C4 7 cm 1 cm phase measurement receiver; L1 only * Note: Selective Availability degradation 5 mm Survey receiver w. post processing (C4) of SPS at ~60 m. level until 1 May 2000 meters in Accuracy Position 1 mm carrier f meas’dat L1 & L2 for iono-correction

Future SPS Systems 100 m Note: IOC for M-codes and C/A-codes on L2 ~08+; current plans call for L5 signals IOC at 2012+ (Using New Civil GPS Signals) 10 m Future 1 Stand-alone SPS receiver (Future 1) C/A-codes (1.023 Mcps) on F1 3.0 m C/A- codes on L1, L2 1 m F2 2.0 m L1 (1575.42 MHz) and Precision SPS stand-alone receiver (F2) L2 (1227.6 MHz C/A-codes on L1, L2 and I,Q -codes on L5 10 cm L5-codes (I5,Q5 at 10.23 Mcps) Real Time Kinematic receiver (F3); carrier φ F3 3 cm on L5 (1176.45 MHz) 1 cm C/A-codes on L1, L2; I5, Q5-codes on L5 Note: L5-codes: I5 on L5 with data msg. F4 5 mm Survey receiver/post process’g (F4); carr. φ Q5 (quad. signal) on L5, no data msg. Position Accuracy in meters in Accuracy Position 1 mm C/A-codes on L1, L2 ; I5, Q5-codes on L5

Current and Future 100 m Note: Current plans call for M-codes IOC at ~2008 Military Systems (PPS) 2 m 10 m 6 m M 2 Military PLGR stand-alone receiver (Mil. 1) Military 1 5 m (Using existing or new signals) 2 - 4 m. C/A and P/Y- codes at L1 only 1 m 0.8 m Military (2f) stand-alone receiver (M2) C/A-codes on L1 (& L2) and 2 m M 3 PPS: C/A +P/Y or M-codes; on L1 & L2 P/Y-codes on L1 and L2 10 cm 50 cm and/or (evolving to) Military DGPS receiver (M3) M-codes on L1 and L2 PPS: C/A +P/Y or M-codes; on L1 & L2 1 cm Position Accuracy in meters in Accuracy Position Note: M-codes are the new Military codes Note: SPS = Standard Positioning Service planned as L1, L2 split spectrum signals 1 mm PPS = Precise Positioning Service

Fig. 9 Position Accuracy Estimates for Civil and Military GPS Receivers in Various Modes of Operation for the 2000 to ~2010+ Time Period Tab. 3 GPS Performance in Various Modes of Operation - for 2000 and 2010

Accuracy estimates for 95% confidence in horizontal; vertical accuracy is about 1.4 x horizontal dimension

Civil C/A-codes at L1 and L2 GPS Bands Position Velocity Time Civil I5, Q5-codes (10.23Mcps) at L5 Diff. Military M-codes P/Y-codes - RNSS - ARNS SA 2000 2000 2000 Comments Mode of Operation L1 L2 L5 GPS 2010 2010 2010 Conventional civil stand-alone Off in 5-10 m 15-30 cm/s 40-100 ns Iono dependent SPS: C/A-code 2000 3-8 m 10-20 cm/s 20-40 ns C/A-code differential 1-3 m 10-20 cm/s 30-60 ns Standard Positioning Service (SPS) 20 cm-1 m 3-10 cm/s 20-30 ns Real time kinematic (RTK) 10-50 cm 5-10 cm/s -- SPS: C/A-code, carrier phase meas. 3-10 cm 1-5 cm/s 5-15 ns Survey: Post-processing; long b 0.5-10 cm L2 carr φ in 00

Civil Current Systems CivilCurrent NA NA SPS: C/A & carrier phase (with 2f) 0.1-3 cm Baseline (b) dep. Conventional civil stand-alone Off in NA NA NA No L2c in 2000 SPS: C/A-codes at L1 and L2 2000 2-5 m 10-20 cm/s 40 ns NA NA NA Code differential *SPS: C/A-codes at L1 and L2 30 cm-1 m 5-10 cm/s 20 ns NA NA NA Precision stand-alone w 3f’s *SPS: C/A & L5 I,Q-codes 1 m-3 m 2-10 cm/s 10 ns

Signals NA NA NA Real time kinematic (RTK); 3f *SPS: C/A, L5 I,Q-codes, carrier φ 1-10 cm 0.5-3 cm/s 2-10 ns

Civil Future (2010+) CivilFuture 1 mradian Precision attitude meas.; 3f’s NA NA *SPS: C/A, L5 I,Q-codes, carrier φ 0.1 mrad Attitude, angle θ Military receiver (1f) Off in 4-8 m 10 cm/s 100 ns E.g., PLGR (P/Y) Precision Positioning Svc. (PPS) 2000 3-6 m 5 cm/s 40 ns Iono dependent 2-4 m 80 ns Military receiver (2f) 10 cm/s Std. 2f rec:(P/Y) PPS: C/A+P/Y or Mil (M) codes 0.5-1 m 5 cm/s 25 ns Future 2f rec:(M) Military DGPS receiver 1-2 m 5-10 cm/s 50 ns

Military Current PPS: C/A+P/Y or M-codes 20-80 cm 1-5 cm/s 10 ns and Future (2010+)

autonomously (without the use of ground-based data) the validity, or integrity, of the navigation information 6 Integrity, Availability, Constellation Size and Power received. In particular, this involves establishing the Concerns credibility of the GPS user equipment measurements and determining if any of the spacecraft are causing an error, System Integrity or are out of tolerance, to the extent that they might provide hazardous or misleading information. Another civil application of concern is the monitoring of the data obtained by GPS airborne receivers to determine

McDonald: The Modernization of GPS 13

Although ranges to four spacecraft are sufficient to compensate for the higher levels of interference and noise determine position, as many as six or more spacecraft in this band. A additional increase of 3-6 dB in the power ranges may be required to determine signal integrity to levels of the civil signals has been promoted by many for the acceptable high confidence level needed to meet a variety of safety, cost and performance reasons. This aviation integrity safety requirements. This fault detection now appears to be planned for the new generations of and elimination (FDE) process, frequently termed GPS spacecraft, or at least for the Block III spacecraft. receiver autonomous integrity monitoring (RAIM), is an The military signals at L1 and L2 are planned for especially difficult problem. The removal of SA improves transmission at higher power levels (by 6-10 dB) than the viability of the FDE process very significantly. Prior current levels. Further, a substantial increase in power to its removal, SA could result in position errors of over beyond this is desired for moderate operational intervals 100 meters for five percent of the time according to its and in selected tactical areas. This would both improve specification. performance in an electronic countermeasures (ECM) Signal Availability environment and provide additional signal robustness. This capability may require a collapsible large aperture The baseline GPS constellation of 24 spacecraft was antenna system. This would provide a steerable spot established for military applications. Most observers beam to cover selected tactical areas. The military needs consider the availability of the spacecraft signals for tend to absorb the substantial prime power requirements some civil applications, especially those relating to of the new spacecraft and other planned functions may safety-of-life applications, marginal or in some cases have some difficulties in the competition. Civil unacceptable. Also, in limited visibility conditions, such improvements may be at risk. as the “urban canyon” situation where a GPS receiver is near ground level with tall buildings on either side, the Figure 10 entitled GPS Signal Power Spectra shows the number of spacecraft in view can be reduced relative signal power spectra for the various existing and

substantially. At times, this results in an insufficient planned GPS signals to be transmitted by the spacecraft.

number of signals to provide a navigation solution. For -60

these and other reasons, there has been interest in C/A

-65 Y

increasing or augmenting the GPS constellation, possibly BOC(10,5)

by 6 to 12 spacecraft, to provide a total of 30 to 36 -70

operational spacecraft. -75

Increased Constellation Size -80

As mentioned, proposals have been made to add GPS (or -85

similar) spacecraft as needed to provide the additional Power Spectrum (dBW/Hz)

robustness desired for the GPS constellation. -90

Investigations accomplished for the FAA and others -95

indicate that 30 to 36 spacecraft may be necessary to -100 -10 -8 -6 -4 -2 0 2 4 6 8 10 meet integrity and other safety-of-life requirements. To Frequency (MHz) date, however, the high cost of placing payloads into orbit has precluded serious consideration of a substantial Fig. 10 Power Spectrum for the GPS M-Codes, P/Y-Codes, C/A- increase in the GPS constellation size. As discussed Codes, C-Codes and L5 I- and Q-Codes earlier, augmentations have been planned to provide Implementation Schedule and Timeliness additional signal availability as well as the differential correction, integrity and ionospheric correction data A principal concern in the modernization and system needed in marine and aviation applications. performance improvement area is the ability of the executive and legislative parts of the government Power Level Improvements organizations involved to recognize the importance Since the modernized GPS spacecraft will provide the associated with GPS modernization, as well as the risk civil community with a C-code signal on the L2 involved in delaying or discarding various aspects of the frequency as a second primary signal, the power program. requirements for this signal are comparable to those of Planned Schedule for GPS S/C Launches the C/A-code signal on L1 (-160dBw). Also, for similar reasons, the P/Y-code power on L2 requires about a four- An example of this concern is the currently planned fold increase (6 dB) in all of the modernized GPS schedule for GPS spacecraft launches. From the spacecraft since the L2 signal was previously used only announced launch information provided by the GPS Joint for ionospheric delay correction. The new L5 signal in Program Office and other sources, Figure 11 was the ARNS band will require a power level about 6 dB prepared, entitled Modernized GPS Capability Dates for higher than that of the C/A-code signal on L1 to Planned GPS Spacecraft Phase-in. This indicates that, at

14 Journal of Global Positioning Systems

GPS Frequency Bands and Phase-in / Phase-out of GPS Spacecraft (All Blocks) Estimates based GPS upon published Space Signal Codes within Bands* Calendar Year S/C lifetimes L1 (1575.42) L2 (1227.6) L5 (1176.45) Vehicle (SV) 1990 2000 2010 2020 2030 Block No’s C/A P/Y M C P/Y M I5 Q5 II, IIA 26 SVs 97 1 - 28 95 Optimistic (28 SVs) ? ? ? SV decay Current capab. II, IIA projection SV-14 Lifetm: 7.6 yrs. IIR 1 - 9 89 05 Existing P/Y- and C/A-codes (8 SVs, 1 Fail) ? ? ? Current capab. 8 SVs Lifetm: 10 yrs. IIR 1-9 97 IIR-M L5 desired but 10 - 21 +M-codes and L2 CS-codes** not planned on IIR-Mod’s S/C modified for 12 SVs (12 SVs) ? ? ? ? ? ? Military M-codes Modified IIR’s have a C-codes on L2** Lifetm: 10 yrs. power limitation IIR-M 10-21 IIF-Lite +L5; I & 03 IOC FOC 1 - 12 Q-codes S/C modified for 2008 2010 (6 SVs firm & Military M-codes ? ? ? 12 SVs Optimistic 6 SVs planned) ? ? ? ? ? C-codes on L2** SV decay Modified Civil L5 signals IIF-Lite 1 - 12 projection Life:12-15 yrs. 05 III 1 - ? ? ? New SVs: GPS III - All signals 18 SVs (IOC) + SV quan. TBD 9-12 IIR-M’s ; 6 IIF-Lites FOC 24 SVs Acquisition 2014 ? ? ? ? ? ? ? ? IOC III 1 - ? in planning S/C built for 2012 Capab’s to 2030 All signals 18 SVs (IOC) ? ? ? ? Life: ~15 yrs. 09 12 IIF-Lites and 6 III’s 1990 2000 2010 2020 2030 * C/A: The C/A-codes, or coarse/acquisition codes (at 1.023Mcps), used currently by both military and civil receivers. ** or alternative civil C signals P/Y: The P/Y-codes, or precise/secure codes (at 10.23Mcps), used by military and other authorized users. M: The M-codes are the new military split spectrum signals with codes of 5.11Mcps at the P/Y-code nulls codes of L1 and L2. L1: The GPS band centered at 1575.42 MHz. L2: The GPS band centered at 1227.6 MHz. L5: The planned civil GPS band centered at 1176.45 MHz. I5: A PRN code (at 10.23 Mcps, 1 ms length, w/data message) for use by the civil community. Signals are in the new civil L5 frequency band . Q5: A PRN code (at 10.23 Mcps, 1 ms, no data) in phase quadrature to I5 on L5. I5 and Q5 are defined by RTCA SC-159 Working Group-1(6/2000). 4.30.99; Rev. a) 1.14.00; b) 8.29.00 c) 11.4.00; d) 2.28.01 e) 6.01 Fig. 11 GPS Modernization: Estimated Availability Dates for Planned Spacecraft (IIR, IIF, III) the current launch rates for planned GPS Block IIR and The European Commission (EC) has sponsored IIF spacecraft, the new modernized capabilities may not investigations of the introduction of satellite systems over be available to civil and military GPS users until about Europe during the past several years. These indicate a 2015-1018. It is clear that significant modifications to the world market for GNSS equipment and services of about GPS spacecraft are needed soon to move the availability $40B in a few years. date for modernized capabilities to an earlier time. This On February 10, 1999, the European Commission (EC) date could be moved to 2010, as shown in the figure, or requested the governments of the 15 states in the possibly earlier, if appropriate modifications (the new European Union (EU) to give their political and financial civil and military signals) to the IIR and IIF spacecraft backing to develop a state-of-the-art Global Navigation can be implemented and if an accelerated schedule for the Satellite System (GNSS), called the Galileo project (EC, production and launch of these spacecraft can be 1999). The EC appears convinced that the development arranged. of a European system would avoid the problems now caused by their dependence on the US GPS and provide substantial economic benefits to the European 7 International Implications community. European Concerns and Galileo Since navigation satellite signals are currently directing air and sea movements, the EC states that there is a clear For some time, the European community (and other market potential for routing cars and trucks more nations worldwide) has had concerns about committing efficiently as well as for farming, fishing, timing and their navigation services to GPS. Their concerns are first, synchronization; and for infrastructure planning, mineral they have no control or role in GPS operation; second, exploration and land surveys. there are no assurances that GPS civil signals will be available in times of conflict involving the US, and third, Signal and User Equipment Compatibility Europe would like to maintain its technological There have been proposals in the US (McDonald, 1998c) competence and participate economically in the and Europe for independently fielded navigation satellite navigation satellite field. The first concern reflects their systems (e.g., GPS and a GNSS) that would use the same strong interest in the commercial, strategic and political civil frequencies and signal structures, as a common base. implications of having European control of the If successful international coordination on this principle positioning, navigation and precision timing services in could be achieved, then the signals from the independent the European region. systems would be compatible and allow worldwide equipment inter-operability. The US DOT has

McDonald: The Modernization of GPS 15 acknowledged the potential benefits of a European GNSS orbital arrangement is not fully defined but current system that is compatible with GPS. Also, the Russian planning indicates the use of spacecraft in medium Earth Federation has indicated its willingness to consider orbit somewhat above GPS that will provide good participating in a joint approach. Russia has access to a coverage to Europe, especially to the high latitude regions set of frequencies (for Glonass) that provide some of the European continent. desirable features for use in a global navigation satellite Combined Performance system. The EC appears willing to proceed constructively on an international basis to develop Galileo and has stated Table 4 illustrates the significant advantages of having that failure for the EU to act now means "missing a huge both GPS and Galileo available to users worldwide. and probably unrepeatable opportunity". Substantial benefits are obtained including improvements in system accuracy, autonomous integrity, interference Cost and Funding mitigation, urban canyon operation and in kinematic The EU estimates the cost of the GALILEO project at precision measurements. These advantages are tabulated between $2-3B, which would be funded in part from the in greater detail on the figure with brief notes provided. EU budget and in part from private and national The improvement in quality of the combined system government sources. The Galileo Project may also operation is impressive. provide a significant role for the Russian Federation. The

Tab. 4 The GPS-Galileo GNSS Characteristics, Features and Combined System Capabilities for existing and Planned Civil Operational Systems

GPS Galileo Combined Characteristic + WAAS S/C + EGNOS S/C Capability Notes

Spacecraft in Orbit 28 + 3 30 + 3 58 + 6 User view: 14–25

Spacecraft Availability (aver.) 8 – 9 8 – 9+ 16 – 18 Excellent,+geom.

Integrity (autonomous) Fair Fair Excellent FDE/ RAIM oper. Coverage (worldwide) Good Good Excellent Espec. high lat's (nominal HDOP-VDOP) Dilution of Precision 1 – 3 1 – 3 0.7 – 2 Improved accur. Interference Susceptibility Low Low Very Low Adaptive f select Safety Services Protection 2 signals 4+ signals 6+ signals Lg. range of sig's Frequencies Available (civil) 1 - 3 1 - 5 2 – 8 User flexibility E911, Related Capabilities Fair Fair Very Good Also greater pwr.

Receiver Cost (relative) 1C 1C 1.2C- Production fcn. - - Accuracy (Autonomous, code)* 1-2 m. 1-2 m. 0.6-1.3 m. Multi-f, estimated * See included charts for accuracy estimates for all classes of GNSS services Figure 12, entitled Comparison of GPS and Galileo power levels, more extensive ground tracking and more Programs illustrates a number of the system frequent and accurate spacecraft position updates. All of improvement considerations relating to GPS and Galileo these dramatically improve accuracy, integrity and other and how they compare in some of the principal areas of aspects of system performance. concern. The current delay in the implementation of GPS Additionally, the management of GPS has changed, and modernization improvements indicates that the Galileo now involves coordinated civil and military funding and program could become the future navigation satellite oversight. Hopefully, other institutional changes will system of choice for a period of several years. Urgent and occur to provide a centralized, coordinated management dedicated action needs to be continued on behalf of GPS. for GPS that will allow the program to compete successfully for future federal funding. 8 Conclusions These factors, combined with the increasing worldwide importance of navigation systems and services, provide a strong basis for integrating GPS into an international 8.1 The Impact of GPS Modernization Global Navigation Satellite System consisting of a number of independent requirement but coordinated The modernization of GPS will provide many new system elements. The modernized GPS will continue to capabilities, such as two new civil frequency bands (L2 play an important military role, as well as a central role in and L5), new civil and military signals, higher signal providing position, velocity, attitude and time services for

16 Journal of Global Positioning Systems

System Improvement Consideration GPS Galileo Primarily DoD management, European Union Centralized management and funding funding and priorities centralized Some IGEB involvement management Guaranteed levels of civil service: Now 1 0 Later 2 2-3 None Autonomous capabilities – Now ... L1(C/A): 5-10 m. After implementation (system/enhancements) L1 (C/A), L2(C); L5: 0.3-3 m. 2-5 Frequencies: 0.3- 3 m.

2012 - 2014+ (~2016)* 2007- 2008+ (~2010)* Schedule for new user full capabilities Assumes launch on demand Poss 5-7yr safety svc lead Civil capabilities in next 10 years Current capabilities; Improved operational (from 1/02, by announced plans, policies) some spacecraft improved performance 2B Euros Funding commitment (current) ~ $1B+ +1- 2B Euros to complete Increased civil signal operational power On some spacecraft Yes levels by 2010? (Rel. to current GPS) (by ~6 dB) (by ~6 dB)

Note: SA removed 1 May 2000 * Indicates normal (realistic) delay of schedule 135.2-3/122.2-5B 99.6, Rev a:00.1.14; Rev.b:00.10.6 Fig. 12 A Comparison of Civil GPS and Galileo Programs - The Galileo Opportunity civil safety, security, science, engineering and related progressed from a 50-60 m. accuracy (at a 95% applications in an economically sensible manner. confidence level) with SA on to a current accuracy (without SA) on an unaided basis of about 5-10 m. In the next decade, GPS modernization will significantly increase the navigation capabilities of both civil and Reduction in Systematic Error Sources includes not only military users by incorporating the following SA removal and ionospheric error correction, but improvements: substantial improvements in GPS receivers, in the control segment redundancy (with the added NIMA monitor New civil frequencies at L2 (1227.6 MHz) and at L5 stations), and improved statistical estimation techniques (1176.45 MHz), as well as retention of the long-standing providing substantially better capabilities for minimizing civil signal at L1 (1575.42 MHz). Military capabilities spacecraft position prediction (ephemeris) errors. will remain in the L1 and L2 bands. The new arrangement provides capabilities to civil users for Increased Signal Availability and Power from GPS ionospheric correction, improved signal robustness, Spacecraft which have greater reliability and lifetimes. increased interference rejection and improved dynamic Power in the new civil L2 signal is to be consistent with precision through the use of techniques for resolving the the current L1 civil signal for greater system robustness. ambiguities associated with precision carrier phase Power in the military M-code signals is to be flexible and measurements. substantially greater than the current P/Y-code power levels. While a larger GPS spacecraft constellation cannot New signal structures for both civil and military users. be guaranteed, there is strong interest in this expansion. The new civil signals at L5 are projected to support a code rate 10 times that of the C/A-code. This will Improved Performance with Augmentations, such as the improve code measurement accuracy, reduce code noise, very substantial performance enhancement achieved by reduce cross-correlation concerns, and provide improved the removal of Selective Availability degradation, as well multipath mitigation. The new military signal structure as the planned augmentations. These include the USCG will provide improved code accuracy, desirable power Differential Network, the National DGPS, the FAA's distribution in the spectrum, and direct access to the WAAS and LAAS, the European EGNOS, the Japanese military secure codes. MSAS, and a large number of other DGPS systems that can provide highly precise position, velocity and time Removal of Selective Availability, which had been measurements for a great variety of applications. planned for between 2000 and 2006 (by the March, 1996 Presidential Decision Directive on GPS), has been International Implications - The GPS has become the de accomplished. GPS now provides full, undegraded facto standard for navigation satellite system accuracy to the civil signals. This, with the additional performance, but there have been long-standing concerns civil frequencies (for ionospheric correction), will internationally because of the US military origin and improve civil GPS performance by a factor of about ten control of the system. However, systematic and (compared to that with SA). For example, GPS has institutional changes have occurred, and are occurring

McDonald: The Modernization of GPS 17 such that GPS now has some new features and some by Vice President Albert Gore and Secretary of excellent opportunities. GPS now has or provides: a) a Transportation Federico Pena, (Document DOT 62-96), joint civil/military management structure, b) an important Executive Office Building, Washington, DC, March 29, national resource with worldwide applications and 1996. implications, c) independent civil and military McDonald, Keith D. (1998) Technology, Implementation and capabilities, both of which are being significantly Policy Issues for the Modernization of GPS and its Role improved, d) a substantial economic “engine” for U.S. in a GNSS, The Journal of Navigation, Vol. 51, No. 3, industry and worldwide users, and e) the capabilities for The Royal Institute of Navigation, Cambridge University taking a leading role, and substantially influencing the Press, The Royal Geographical Society, 1 Kensington Gore, London SW7 2AT, September 1998. formation of an international GNSS. McDonald, Keith D. (1998) The GPS Modernization Dilemma and Some Topics for Resolution, The Quarterly 8.2 Final Comment Newsletter of he Institute of Navigation, Vol. 8, No. 2, Summer, 1998. For the modernization of GPS to occur in a timely and McDonald, Keith D. (1998c) The Modernization Mantra, GPS useful manner. it is clear that strong funding support, a World, Directions '99, Vol. 9, No. 12, p.46, Advanstar strong national commitment, possibly White House Communications, 859 Willamette St., Eugene, OR 97401, leadership, new institutional arrangements and a variety December, 1998. of other factors are necessary. The technical and NAPA (1995) The Global Positioning System - Charting the implementation issues appear straightforward; however, Future, A Joint Report of the National Academy of Public the institutional, funding, international and national Administration (NAPA) Panel and the National Research priority concerns appear critical. The modernized GPS Council (NRC) Committee, Document available from with the introduction of the European Galileo system and NAPA, Washington, D.C., 1995. augmentations to both provides a tremendous combined NRC (1995) The Global Positioning System - A Shared capability that can benefit civil users worldwide. National Asset; Recommendations for Technical Improvements and Enhancements, Committee on the Future of the Global Positioning System, Commission on References Engineering and Technical Systems, National Research Council, National Academy Press, Washington, D.C., 1995. EC (1999) Galileo to Set Pace in Satellite Navigation, Improve Safety, Generate Jobs, Commission Says, European Spilker, James J., et al. (1998) A Family of Split Spectrum GPS Commission Public Information Release, Publication Civil Signals, Proceedings of the 11th International IP/99/102, European Commission, Brussels, 10 February Technical Meeting of the Satellite Division of the Institute 1999. of Navigation, Part 2, ION GPS-98, The Institute of Navigation, 1800 Diagonal Rd., Alexandria, VA 22314, Gore Albert (1999) Vice President of the United States, New September 15-18. Global Positioning System Modernization Initiative, The White House, Office of the Vice President, Contact (202) USGAO (1998) The U.S. General Accounting Office, A Study 456-7035, Public Announcement on the Global Positioning of the Market Potential and Economic Factors System, Washington, DC, January 25, 1999. Influencing the Global Positioning System, US Gov. Printing Office. Gore Albert and Pena Federico (1996) Presidential Decision Directive on the Global Positioning System, Announced

Journal of Global Positioning Systems (2002) Vol. 1, No. 1: 18-24

Precise Ionosphere Modeling Using Regional GPS Network Data

Y. Gao and Z.Z. Liu Department of Geomatics Engineering The University of Calgary Calgary, Alberta, Canada T2N 1N4 e-mail: [email protected]; Tel: 403-220-6174; Fax: 403-284-1980

Received: 2 July 2002 / Accepted: 18 July 2002

Abstract. The ionosphere affects the electromagnetic must be estimated so that a correction can be made to waves that pass through it by inducing an additional eliminate it from the GPS observations. Precise transmission time delay. The ionosphere influence has ionosphere effect estimates are also important for space now become the largest error source in GPS positioning weather research and earth observation applications and navigation after the turn-off of the Selective (Komjathy, 1997). Availability (SA). In this paper, methods of 2D grid- The currently available ionosphere correction models based and 3D tomography-based ionospheric modeling include the Klobuchar ionosphere parameters broadcast are developed based on regional GPS reference networks. from the GPS satellites but the Klobuchar model could Performance analysis was conducted using data from two only correct about 50% of the total ionosphere effects different regional GPS reference networks. The modeling (Klobuchar, 1987). More precise ionosphere model is accuracy of the vertical TEC (VTEC) is at the level of therefore required. Ionosphere modeling methods using several TECU for 2D ionospheric modeling and about GPS data from ground GPS networks have been one TECU for 3D tomographic modeling after a extensively investigated in the past several years comparison to independent ionospheric map data or (Komjathy, 1997; Skone, 1998; Jakowski et al., 1998; directly measured ionospsheric TEC values. The data Liao, 2000; Fedrizzi et al., 2001). Komjathy (1997) analysis has indicated that the modeling accuracy based established a polynomial ionospheric model based on on the 3D tomography method is much higher than the data from reference stations of International GPS Service 2D grid-based approach. (IGS) and has compared it to the TOPEX/Poseidon- derived (T/P) TEC data. Agreements at the level of 5 Key words: GPS, Ionosphere, Ionosphere Grid, TECU were reported under medium and low solar Tomography. activity conditions (Komjathy, 1997). Skone (1998) employed a two-dimensional grid-based model to characterize the ionosphere activities over the auroral region and an accuracy of about 34 cm was obtained. In Gao et al. (2002), ionosphere parameters are estimated along with satellite and receiver biases using data from a 1 Introduction regional area GPS network. To date, all proposed ionospheric models could be classified into two different The ionosphere affects the electromagnetic waves that categories: grid-based and function-based. The early pass through it by inducing an additional transmission modeling methods are mostly based on the function time delay. The magnitude of this effect is determined by fitting techniques such as the broadcast ionosphere model the amount of total electron content (TEC) and the from the GPS satellites (Klobuchar, 1987), the frequency of electromagnetic waves. Under normal solar polynomial functions (Coster et al., 1992; Komjathy, activity conditions, this influence on GPS signals is 1997) and the spherical harmonics (Schaer, 1999; usually in the range from a few meters to tens of meters Walker, 1989). On the other hand, the grid-based method, but it could reach more than 100 meters during severe first proposed by the MITRE Corporation and the Air ionosphere storms. After the turn-off of the Selective Force Phillips Laboratory (El-Arini et al., 1994 and Availability (SA), the ionosphere effect has become the 1995), has demonstrated its capability for higher largest error source in GPS positioning and navigation. modeling accuracy when compared to the function-based For high precision GPS positioning, the ionosphere effect algorithms. The grid-based modeling technique has since

Gao et al:: Precise Ionosphere Modeling Using Regional GPS Network Data 19 been extensively used for both global and regional ρ )( −++++−+= BbIkdddTdtcP 2 orb trop 1 2 PP 2 network-based ionosphere recovery (Gao et al., 1994; d ++ ε P )( (3) FAA, 1997; Skone, 1998; Liao, 2000). / Pmult 2 2 No matter whether it is function or grid-based, current ρ )( λ +−+++−+=Φ bIkddNdTdtc 2 22 orb trop 1 Φ2 ionosphere modeling is two-dimensional in nature, which dB ε Φ++− )( (4) assumes that the ionosphere is condensed on a single Φ2 mult / Φ2 2 shell at a fixed altitude above the earth surface. This where assumption, however, is only an approximation to the reality and it is not physically true. In order to further 2 2 2 fk ii /(f1 f 2 ), =−= 1,i ;2 improve the ionosphere modeling accuracy and the ρ model’s sensitivity to temporal ionosphere variations, is the true geometric range between receiver and ionospheric tomography modeling method has started to satellite (m); receive more attentions in the recent years (Raymund et c is the speed of light (m/s); al., 1990 and 1994; Raymund, 1995; Howe, 1997; Liu et dt is the satellite clock error with respect to GPS al., 2001a and 2001b). An ionosphere tomographic model time (s); can describe the ionosphere field in a three-dimensional dT is the receiver clock error with respect to GPS frame instead of a two-dimensional frame as used by time (s); previous methods. As a result, the ionosphere L tomography method would allow for more precise λi is the wavelength of GPS signal on i (m); exploration of the ionospheric characteristics and Ni is the carrier phase integer ambiguity (cycle); subsequently for more precise modeling accuracy. d trop is the tropospheric delay (m); This paper describes the recent research results in the area I is the ionospheric delay parameter (m); of high precision ionosphere modeling using regional d GPS reference network data and focuses on 2D grid- orb is the satellite orbit error (m); based and 3D tomography-based ionospheric modeling. d The paper is organized as follows. In Section 2, mult is the multipath effect (m); ionospheric delays are derived from GPS dual-frequency b is the satellite hardware delay (m); observations including an algorithm for carrier phase B is the receiver hardware delay (m); and leveling on code-derived ionospheric delay ε( ) is the measurement noise (m). measurements. Section 3 discusses a 2D ionosphere modeling method while a 3D ionosphere tomographic Differencing the code observations from L1 and L2 model is presented in Sections 4. Numerical results and results in the following ionosphere measurements: performance analysis are provided in Section 5. ε −+−+−=− PPBbIPP )( (5) Conclusions are given in Section 6. 21 21 where −= bbb , −= BBB . b and B represent 1 PP 2 1 PP 2 the differential hardware delays between the L1 and L2 2 GPS IONOSPHERE MEASUREMENTS frequencies and they are often referred to as satellite and receiver L1/L2 inter-frequency biases. Although these A dual-frequency GPS receiver used at a reference station biases are actually time dependent, in practice they are consists of both code and carrier phase observations on very stable over time on a scale of days to months so that L1 (1575.42 MHz) and L2 (1227.60 MHz) frequencies, they can be treated as constants during ionosphere denoted as Pi and Φi (i = 1, 2) in the following. modeling (Gao, et al, 1994; Schaer, 1999). As to the bias Mathematically the corresponding observations can be magnitude, the satellite inter-frequency bias is usually in described as the range of several nanoseconds while the receiver inter- frequency bias could be as large as more than 10 L Frequency: 1 nanoseconds (Gao et al., 1994). ρ )( −++++−+= BbIkdddTdtcP 1 orb trop 2 1 PP 1 Considering the much higher noise level of the

d / Pmult ++ ε P1 )( (1) ionosphere measurements derived from the code 1 measurements, the carrier phase observations from L1

1 ρ )( λ 11 orb trop 2 +−+++−+=Φ bIkddNdTdtc Φ and L2 described in equations (2) and (4) can be used to 1 smooth the code observation for a more precise vertical dB ε Φ++− )( (2) Φ1 mult / Φ1 1 TEC estimate. Such carrier phase smoothing technique is also often referred as “carrier phase leveling”. Given L Frequency: 2 below is a smoothing function that has been described in Gao et al. (2002).

20 Journal of Global Positioning Systems

______w )( w )( An ionospheric grid model consists of grids distributed 1 k 2 k o o PP 21 )( k =− PP 21 )( k +− on the ionospheric shell in preset spacing usually 3 x 3 . + ww )()( + ww )()( 1 k 2 k 1 k 2 k Any pierce point therefore will fall within a specific grid ______defined by its surrounding four grid points. The slant PP δ Φ−Φ+− ])()([ (6) 21 k−1 kk −1,21 ionospheric delay at a pierce point, which has elevation ______angle of E, can be linked to the vertical ionospheric delay where k is the time epoch index; − PP 21 )( is the at the same location by a mapping function sf(E). smoothed ionosphere measurement and Meanwhile the vertical ionospheric delay at the pierce point can be described by the vertical ionospheric delays 1 w )( = (7) at its surrounding grid points. The relationship between 1 k 2 the slant ionospheric delay at the pierce point and the σ −PP )( 21 k vertical ionospheric delays of its surrounding four grid 1 points can be expressed by the following equation: w )( = 2 k 2 2 (8) + σσ 4 −PP 21 )( k δ Φ−Φ 21 )( k ______21 =− )()( ∑ )( ⋅ pIEsfPP kkv (10) δ kk − 211,21 k Φ−Φ−Φ−Φ=Φ−Φ )()()( k−121 (9) k =1

Note that the smoothed ionosphere measurements in 2 Esf 1/1)( −= [cos( + RhE )]/1/() (11) equation (6) are still corrupted by the inter-frequency biases bP and BP which therefore need to be estimated where along with the ionospheric delay parameters. E is the satellite elevation angle; R is the earth radius; h is the height of the ionosphere shell above the 3 2D IONOSPHERE MODELING earth’s surface; (I )kv is the vertical ionospheric delay parameter at the The ionosphere is a region of ionized plasma that extends grid point k; from roughly 50km to 2000km above the surface of the p is a weighting function which is used to project earth. Generally, the ionosphere can be divided into k the vertical ionospheric delay at grid point k to several layers in altitude according to electron density, the pierce point. which reaches its peak value at about 350km in altitude. For 2D ionospheric modeling, the ionosphere is assumed Taking into account the existence of L1/L2 inter- to be concentrated on a spherical shell of infinitesimal frequency biases b and B, we can establish the following thickness located at the altitude of about 350km above the ionosphere observation equation for a satellite (j) and earth’s surface (Gao et al., 1994). The implementation of receiver (i) pair: the single-layer grid model requires computation of the intersection of the line-of-sight between the GPS receiver ______4 =− )()()( ij −+⋅ BbpIEsfPP (12) and the observed satellite on the ionosphere shell as 21 ij ∑ vij k ijijk illustrated in Figure 1. The intersection point of the GPS k=1 signal with the ionospheric shell is defined as pierce point Equation (12) is the fundamental equation for the grid at which the slant ionospheric delay has an elevation point vertical ionospheric delay estimation in grid-based angle of E. 2D ionospheric modeling using carrier smoothed code- derived ionosphere measurements. As to parameter GPS satellite j Pierce point estimation, the standard least squares or Kalman filtering Ionosphere shell method can be used to facilitate the optimal estimation of the vertical ionospheric delay parameters and the Ionosphere E satellite/receiver inter-frequency biases.

Receiver i e 4 3D Tomographic modeling

Earth surface A two-dimensional (2D) ionospheric model as described h =350km+R R in the last section has difficulty to characterize the ionospheric field in the full spatial dimensions because it Earth center is unable to provide the vertical ionospheric profile. A O three-dimensional (3D) ionospheric model therefore is expected. A 3D ionospheric modeling method using Fig. 1 Ionosphere Shell

Gao et al:: Precise Ionosphere Modeling Using Regional GPS Network Data 21 tomography technique will be described in the following m where Pn φ)(cos is the associated Legendre polynomial which is able to characterize not only the ionospheric of order m and degree n ; (zZ ) is the empirical horizontal profile but also the vertical one. k orthogonal function (EOF); a m and bm are the Tomography-based modeling consists of two nk nk fundamental steps. First, integral measurements are made coefficients to be estimated. Combining Equations (14) of the medium of interest, ideally along many paths from and (15), the observation equation for ionospheric 3D many different viewing angles. Second, these integral modeling can then be established by the following measurements are inverted to obtain an estimate of the equation: field (Howe, 1997). In ionospheric tomography, the sat K M M m m integral measurements are the lines of sight between the TEC = nk λ + nk mbma λ)]sin()cos([ ⋅ ∫ rx ∑∑∑ GPS receivers and satellites, which pass through the k=−1 ==Mm mn entire ionosphere. A 3D ionospheric model can be m φ)(cos ZP )( dsz constructed using the tomography technique, horizontally n k by spherical harmonics functions (SHFs) and vertically K M M sat by the empirical orthogonal functions (EOFs). The m m = ank n φλ k )()(cos)cos( dszZPm ∑∑∑ ∫ rx harmonics functions are primarily formed by the first k=−1 ==Mm mn several orders of harmonics. The high-order harmonics K M M sat however improve the sharpness of the fronts. EOFs are m m + bnk n φλ k )()(cos)sin( dszZPm (16) derived from existing data set (observation data or model ∑∑∑ ∫ rx k=−1 ==Mm mn data) of the medium of our interest. The observation equation of ionospheric tomography can be given as Equation (16) is the fundamental observation equation for follows (Liu et al., 2001a and 2001b): 3D ionospheric modeling using tomography inversion sat sat technique, through which the GPS derived total electron 0 (13) TEC = e φλ = e + e φλδφλ )],,(),,([),,( dszNzNdszN ∫ rx ∫ rx content (TEC) and the coefficients describing the ionosphere field are linked. The rest task of ionospheric where TEC is the total electron contents along the line of tomography is to optimally estimate the model sight from a GPS satellite to a ground GPS receiver and coefficients in equation (16) in which the number of the e φλ zN ),,( is the ionospheric electron density at the unknown model parameters is determined by the 0 truncation limits of SHFs and EOFs. A so-called geospatial position of ( φλ ,, z ). e φλ zN ),,( is the a priori weighted, damped least squares technique has been value of N φλ ,,( z) which could be an output from an e developed via a combination of weighting and damping empirical model that reflects the deterministic portion of procedures described in Liu et al. (2001a and 2001b). our a priori information. δNe ( φλ ,, z) is the correction to the a priori value. λ φ,, z are longitude, latitude and altitude, respectively, referenced to a solar-geomagnetic 5 Results AND ANALYSIS coordinate system. For the convenience purpose, we The data from two regional GPS reference networks, could let N 0 φλ z),,( be approximately equal to zero, e SWEPOS and SCIGN, were processed to generate 0 namely N e φλ z),,( = 0 . Thus we have, regional ionospheric model using the developed 2D grid- based and 3D tomography techniques. sat TEC = e φλδ ),,( dszN (14) ∫ rx 5.1 Results of 2D Grid-Based Modeling The correction for the electron density function e φλδ zN ),,( can be modeled by a series of space-location The data from the Swedish GPS reference network related functions. More specifically we can employ (SWEPOS) has been used to evaluate the performance of spherical harmonic expansions horizontally and empirical the 2D grid-based ionosphere modeling method. The orthogonal functions (EOFs) vertically to model the SWEPOS network (Figure 2) consists of 21 continuously ionospheric electron correction term. The integration of operating GPS stations extending from latitude 55 to 69 these two sets of functions enables us to depict the degrees north with an average station separation of ionosphere field in a 3D mode as follows: around 200km (Liao, 2000). Data used in the numerical K M M analysis includes a total of five days of consecutively m m tracked GPS data collected during May 17-21, 1999. For e φλδ zN ),,( = ∑∑∑ nk λ + nk mbma λ)]sin()cos([ ⋅ k =−1 ==Mm mn the purpose of this research, only ten stations in the southern portion of the SWEPOS (the two stations in the m n φ k zZP )()(cos (15) block frame are not included due to an accident loss of

22 Journal of Global Positioning Systems data) were used in the numerical computation. The ten between UofC and CODE estimates as well as the RMS stations are located within a range of approximately 55oN value over the consecutive five days. The TEC estimates to 60oN in geographic latitude and 11oE to 18oE in in the central region of the ground SWEPOS network geographic longitude. (56oN and 59oN) are very consistent to the estimates derived from the CODE global TEC maps. The overall agreement was at the level 2.4~3.0 TECU. On the other hand, the consistency has been degraded for grids off the central latitude region (53oN and 62oN) to a level of up to 4.6 TECU. This is due to the fact that the ionosphere measurement density in the central region is higher than other boundary regions over a regional network.

40 30 UofC 20 CODE 10

0 Fig. 2 Reference Stations in SWEPOS

Vertical TEC (TECU) 0123456 Shown in Figure 3 and 4 are the vertical TEC estimates at Geomagnetical Local Time (Day) latitudes of 59oN and 56oN respectively with a grid size of 3o × 3o based on the five-day data. The results show clearly the diurnal behavior of the ionosphere and the variations from day to day. These vertical TEC estimates Fig. 3 VTEC at Geomagnetic Latitude 59oN are consistent with diurnal ionosphere behavior considering the overall change pattern. For external comparison purpose, the vertical TEC estimates derived 30 from the global ionosphere maps (GIM) provided by the 20 UofC Center of Orbit Determination in Europe (CODE) are CODE also shown in Figure 3 and 4. CODE is one of the five 10 International GPS Service (IGS) ionosphere analysis 0 centres that have supplied global TEC maps since June 1, Vertical TEC (TECU) 0123456 1998 on a regular basis to the Crustal Dynamics Data Information Systems (CDDIS), a global data centre of Geomagnetical Local Time (Day) IGS. The overall consistency in terms of RMS values between the global TEC maps derived from different analysis centres is reported to be at the level of 3~5 Fig. 4 VTEC at Geomagnetic Latitude 56oN TECU although the inconsistency could reach up to 7~10 TECU between some centres. Provided in Table 1 are the daily RMS values of the TEC estimate differences Tab. 1 RMS Values of TEC Estimate Differences (TECU)

Grid Latitude Day 1 Day 2 Day 3 Day 4 Day 5 Day 1~5

53oN 3.47 5.78 2.50 2.78 3.71 3.83

56oN 1.90 3.60 1.93 1.42 2.34 2.36

59oN 1.89 4.53 3.89 1.64 1.29 2.96

62oN 3.46 6.98 5.79 3.08 1.62 4.61

Gao et al:: Precise Ionosphere Modeling Using Regional GPS Network Data 23

4.0 5.2 Results of 3D Tomographic Modeling 3.0 2.0 The performance analysis of 3D tomographic modeling 1.0 was conducted using GPS data observed from six GPS 0.0 -1.0 reference stations within the Southern California Error (TECU) Integrated GPS Network (SCIGN) on May 15, 2000. The -2.0 SCIGN consists of 250 GPS stations and is primarily 0 100 200 300 400 designed to monitor the crustal deformation and Number of Ionospheric Delays earthquake activities in southern California region (Mean=0.56 TECU, RMS=1.17 TECU) (SCIGN, 2002). The geographical distribution of the selected six stations is shown in Figure 5. A total of 60 Fig. 6 Vertical Ionospheric Tomography Modeling Errors epochs of dual frequency GPS measurements from six TEC −TEC stations at a data rate of 30 seconds were used for the RE = direct model 100% (17) construction of an ionospheric tomography model for the TECdirect region. To assess the modeling accuracy, the 300 measurements, collected immediately after the 60 epochs, The results indicated that the model-derived slant were used as prediction data points for comparison ionospheric delays have a mean error of 1.4% with purpose. respect to the directly measured ionospsheric delay values. Most relative errors varied between –4.0% and +4.0%. The 3D tomographic modeling therefore has produced much lower ionospheric modeling errors or better modeling accuracy than the 2D grid-based modeling method. It is also worth to mention that a “carrier phase leveling” procedure was not implemented during the 3D tomography ionospheric modeling computation. So it is expected that a further accuracy improvement can be obtained if the noise level of the code-derived ionospheric delay measurements is reduced via carrier phase smoothing.

12.0 8.0 4.0 0.0 -4.0 Percent (%) -8.0 0 100 200 300 400 Fig. 5 Distribution of Six SCIGN Reference Stations Number of Ionospheric Delays Once the tomography model coefficients were obtained, (Mean Error=1.40%) the obtained model parameters are used to calculate the ionospheric delay for any line of sight between a GPS Fig. 7 Slant Ionospheric Tomography Modeling Error in Percentage receiver and satellite. If the model derived ionospheric delays can be compared to its directly measured value from the dual frequency GPS observations, the 3D 6 Conclusions tomography ionospheric modeling accuracy can then be assessed. The comparison results are provided in Figure 6 Both two and three-dimensional ionospheric modeling where the differences indicated an agreement for the methods have been described in this paper and their vertical ionospheric delays between the model-derived performance have been assessed based on data from two estimates and the direct measurements at the level of 1.2 different regional GPS reference networks. The TECU. agreement for 2D grid-based modeling is at the level of Shown in Figure 7 are the error estimates in percentage several TECU when compared to an independent source calculated using the following equation: while the agreement for 3D tomographic modeling is at the level of about one TECU compared to the directly measured ionosphere values from dual-frequency GPS

24 Journal of Global Positioning Systems observations. Based on the developed 3D tomographic Howe, B.M. (1997). 4-D Simulations of Ionospheric ionospheric model, the slant ionospheric delays could be Tomography, Proceedings of ION National Technical well recovered with a mean error of 1.4%, indicating a Meeting, Santa Monica, California January 14-16, 1997. significant improvement of accuracy over the 2D grid- Jakowski, N., E.Sardon and S. Schlüter (1998). GPS based TEC based modeling. More numerical analysis, however, is Observations in Comparison with IRI95 and the required to further investigate the performances of the 2D European TEC Model NTCM2, Advances in Space and 3D ionospheric models presented in this paper. Research, Vol. 22, pp.803-806. Klobuchar, J.A. (1987). Ionospheric Time-Delay Algorithm for Single-Frequency GPS Users, IEEE Transactions on Acknowledgements Aerospace and Electronic Systems, Vol. AES-23, No.3, pp.325-331. This research was supported by a strategic grant from the Komjathy, A. (1997). Global Ionospheric Total Electron Natural Sciences and Engineering Research Council of Canada Content Mapping Using the Global Positioning System, (NSERC). The Southern California Integrated GPS Network Ph.D. dissertation, Department of Geodesy and Geomatics and its sponsors, the W.M. Keck Foundation, NASA, NSF, Engineering Technical Report NO. 188, University of New USGS, SCEC, are thanked for providing data used in this study. Brunswick, Fredericton, New Brunswick, Canada, 248pp. Liao, X. (2000). Carrier Phase Based Ionosphere Recovery References Over A Regional Area GPS Network, UCGE Reports, Number 20143, University of Calgary, Calgary, Alberta, Coster, A.J., E.M. Gaposchkin and L.E. Thornton (1992). Real- Canada, pp120. Time Ionospheric Monitoring System Using GPS. Liu, Z.Z. and Y. Gao (2001a). Optimization of Navigation, Journal of Institute of Navigation, Vol.39, Parameterization in Ionospheric Tomography, No.2, pp.191-204. Proceedings of ION GPS 2001, Salt Lake City, Utah, USA, El-Arini, M.B., C.J. Hegarty, J.P. Fernow and J.A. Klobuchar September 11-14, 2001. (1994). Development of an Error Budget for a GPS Wide- Liu, Z.Z. and Y. Gao (2001b). Ionospheric Tomography Using Area Augmentation System (WAAS), Proceedings of The GPS Measurements, Proceedings of KIS-2001, Banff, Institute of Navigation National Technical Meeting, San Alberta, Canada, June 5-8, 2001. Diego, CA, January, 1994. Raymund, T.D. (1995). Comparisons of Several Ionospheric El-Arini, M.B., R.S. Conker, T.W. Albertson, J.K. Reagan, J.A. Tomography Algorithms, Ann. Geophys., 13, pp. 1254- Klobuchar and P.H. Doherty (1995). Comparison of Real- 1262. Time Ionospheric Algorithms for a GPS Wide-Area Augmentation System (WAAS). Navigation, Journal of Raymund, T.D., J.R. Austen, S.J. Franke, C.H. Liu, J.A. The Institute of Navigation, Vol.41, No.4, pp.393-412. Klobuchar and J. Stalker (1990). Application of Computerized Tomography to the Investigation of FAA (1997). Specification for the Wide Area Augmentation Ionospheric Structures, Radio Science, 25, pp. 771-789. System, FAA-E-2892C (draft), 1997. Raymund, T.D., Y. Bresler, D.N. Anderson and R.E. Daniell Fedrizzi, M., R.B. Langley, A. Komjathy, M.C. Santos, E.R. de (1994). Model-assisted Ionospheric Tomography: A New Paula and I.J. Kantor (2001). The Low-Latitude Algorithm, Radio Science, 29, pp. 1493-1512. Ionosphere: Monitoring its Behavior with GPS, Proceedings of ION GPS 2001, 14th International SCIGN (2002). Southern California Integrated GPS Network, st Technical Meeting of the Satellite Division of the Institute http://www.scign.org, accessed on April 1 , 2002. of Navigation, Salt Lake City, UT, September 11-14, 2001. Schaer, S. (1999). Mapping and Predicting the Earth’s Gao, Y., P. Heroux and J. Kouba (1994). Estimation of GPS Ionosphere Using the Global Positioning System, Ph.D Receiver and Satellite L1/L2 Signal Delay Biases Using dissertation, Astronomical Institute, University of Berne, Data from CACS, Proceedings of KIS-94, Banff, Canada, Switzerland, pp205. August 30 - September 2, 1994. Skone, S. (1998). Wide Area Ionosphere Grid Modeling in the Gao, Y., X. Liao and Z.Z. Liu (2002). Ionosphere Modeling Auroral Region. UCGE Reports Number 20123, Ph.D Using Carrier Smoothed Ionosphere Observations from a thesis, The University of Calgary, Calgary, Alberta, Regional GPS Network, Geomatica, Vol. 56, No.2, pp. 97- Canada. 106. Walker, J.K. (1989). Spherical Cap Harmonic Modeling of High Latitude Magnetic Activity and Equivalent Sources with Sparse Observations, Journal of Atmospheric and Terrestrial Physics, Vol.51, No.2, pp.67-80.

Journal of Global Positioning Systems (2002) Vol. 1, No. 1: 25-33

Multipath Mitigation for Bridge Deformation Monitoring

G. W. Roberts, X. Meng, A. H. Dodson, E. Cosser Institute of Engineering Surveying and Space Geodesy, The University of Nottingham, University Park, Nottingham. NG7 2RD. UK e-mail: [email protected]; Tel: +44(0)115-951-3933; Fax: +44(0)115-951-3881

Received: 24 June 2002 / Accepted: 10 July 2002

Abstract. GPS carrier phase multipath with varying deformations and dynamic responses to active loadings, amplitudes of up to several centimetres and periods of such as wind, temperature change, traffic, and even couple of minutes is a major error source, which affects earthquakes (Duff et al. 1998; Ogaja et al. 2001; Roberts the correct interpretation of bridge deformation. In this et al. 1999; Roberts et al. 2000). The advance of data paper, a recursive adaptive filtering (AF) algorithm has processing algorithms further makes detailed component been employed to mitigate multipath signature in the analysis and diagnosis of structural deflection possible coordinate time series of GPS solutions. In order to (Meng 2002). It provides the opportunity for GPS maximise the suppression of the multipath signature, researchers to conduct subtle studies into the potential exact alignment of the input time series into the AF error sources and their impacts on the data quality. system is crucial. An algorithm using the cross- Furthermore, it makes correct error modelling possible correlation coefficient of day-to-day GPS coordinate time from such a component analysis. Of many error sources series is presented to align GPS coordinates. To isolate relevant to satellite geometry and signal propagation multipath from the contaminated GPS coordinate time media, multipath is still a research focus for precise GPS series, relative displacements calculated from the positioning (Rizos 1999). The proposed approaches for accelerations sensed simultaneously with GPS receiver mitigating multipath effects are far from practical and by a triaxial accelerometer housed in a specially designed effective, especially when GPS is applied to monitor cage is used as the reference signal sequence within the dynamic structural deformation. Ge et al. (2002) present a AF system. Associated algorithm for the relative velocity simple approach using an adaptive finite impulse and displacement calculation is also introduced in the response (FIR) filtering technique to reduce the impacts paper. It demonstrates that it is possible to achieve of multipath on continuous GPS (CGPS) sites. millimetre positioning accuracy by the AF approach and Since 1993, the Institute of Engineering Surveying and an integrated sensor system of GPS receiver and triaxial Space Geodesy (IESSG) at the University of Nottingham accelerometer. has initialised studies of the applications of kinematic GPS on structural deformation monitoring (Ashkenazi et Key words: Multipath mitigation, Structural deformation al. 1997). The research emphases in the past were on the monitoring, Adaptive filtering, GPS and accelerometer implementations of a GPS-based monitoring system, data integration, Cross-correlation. collection, and simple analysis. Further work in this area discloses that an integrated system with other sensors is necessary in providing precise measurements and hence a robust monitoring system. It was also recognised that sophisticated signal identification, data processing, and deformation analysis techniques are essential. Recent 1 Introduction research focuses of the IESSG in this field are on sensor integration of triaxial accelerometers with dual frequency With the development of real-time kinematic (RTK) GPS geodetic GPS receivers (Roberts et al. 2001) and even receiver and antenna technology, GPS is currently used in more recently with single frequency geodetic GPS areas where high measurement precision is required receivers. A new multipath mitigation algorithm based on within high dynamic environment. Attempts have been recursive adaptive FIR filtering (AF) in the dynamic made in recent years to investigate the feasibility of environment of bridge deflection monitoring scenario is applying GPS technology to monitor structural

26 Journal of Global Positioning Systems proposed by Dodson et al. (2001) and further studied by environment is the main purpose of applying AF Meng et al. (2001) and Meng (2002). approach in the data processing. In this paper, the fundamental of the AF technique is Figure 1 is a simple schematic of an adaptive system that briefly reviewed. GPS data collected on the turret of the consists of an FIR filter (processor) and an adaptive IESSG building with Leica SR510 single frequency GPS algorithm. d is the application provided input signal or receivers and associated antennas are employed to desired sequence. It can be the real measurement time investigate the day-to-day multipath signature. A cross- series for a specific process that can be compared with correlation algorithm is used to estimate the time shift of the FIR filter predicted output y . x is the reference the day-to-day position time series. The multipath pattern signal which is used to output prediction y with Eq. 1. In isolated from normally time shifted coordinate sequences of two consecutive days (four minutes) is compared with Eq. 1, a reference sequence X including the current the exactly aligned coordinate time series, according to input x with sequence length Q , which is the filter the estimated time shift of each individual direction in length, is employed to estimate instantaneous prediction WGS84 during the two days. It attempts to emphasise the at epoch n . Parameter b is the filter coefficient. importance of aligning the input signals to an AF system. Q−1 As an alternative to mitigate multipath, a simple = ∑ q − qnxbny )()( (1) algorithm is presented in the paper to calculate the q=0 relative displacements from the accelerations sensed by a triaxial accelerometer, which is physically connected with In the actual calculation, it is possible to start with the GPS antenna by a specially designed cage. Spectral arbitrary initial values of filter coefficient sequence v T analysis approach is applied to the input and output time = ,,( 210 ,..., bbbbB Q ) . Then as each new input sample series to evaluate the efficiency of adaptive filtering in nx )( enters the adaptive filter, the corresponding output suppressing multipath in a dynamic environment. Through the above geodetic signal diagnosis and or filter prediction (ny ) is made and compared with analysis, millimetre positioning accuracy can be achieved nd )( . The error signal e(n −= ()() nynd ) is formed, and which further confirms that with the proposed procedure used to update the filter coefficients based on the method even in a multipath hostile environment, an integrated of steepest descent using Eq. 2 (Haykin 2001). µ is a sensor system of GPS receivers and accelerometers can parameter that controls the rate of convergence of an AF still provide robust and highly accurate positioning approach. The updated filter coefficients are then solution. employed in the AF processor to predict the next signal output with the coming of new reference signal by an adaptive algorithm. 2 Alignment of Day-to-day GPS Position Solutions vv r +1 nn += µ )(2 XneBB n (2)

2.1 Fundamental of Adaptive Filtering (AF)

Multipath, receiver random noise, and unmodelled relative tropospheric delay are major error sources under the scenario of a bridge deformation monitoring as analysed by Meng (2002). Multipath impact on the positioning solution is related to the repetition of the satellite constellation between two sidereal days. This characteristic can be exploited to extract multipath signature from the positioning time series. If closely setup reference stations are employed and the height differences between the stations are within several tens of metres, the unmodelled tropospheric delay can be reasonably neglected. Otherwise meteorological data Fig. 1 Schematic of an AF system should be collected to cope with these residuals. Receiver random noise characterises pure white noise if systematic In the application of AF approach, filter length and errors are properly modelled. So, in this case, three convergence parameter are two factors that affect the potential components form the positioning solution performance of an AF system. The determinations of the sequence, which are the real bridge movement, receiver optimal values of these parameters are discussed by random noise, and multipath signature. Isolation of these Meng (2002). components in a dynamic bridge deformation

Roberts: Multipath Mitigation for Bridge Deformation Monitoring 27

2.2 Result Comparison of Different Alignment Since it is very difficult to correct the multipath impact Approaches directly from the carrier phase measurements by using day-to-day approach, it limits the applications of GPS in As explained by Roberts et al. (2002), finding the exact many areas where high accurate positioning solutions are match point of two time series is very important in order required in a real-time mode. Therefore, the final to mitigate multipath and isolate deformation effectively. positioning solutions via post-processing of raw measurements are usually used to estimate the time shifts

Suppose two time series di and xi , which could be the in three dimensions in a specific coordinate system for coordinate time series at one observation site on two mitigating coordinate multipath. consecutive days, or the raw pseudorange measurements. To evaluate the performance of Leica single frequency di and xi could be same length vectors or vectors of SR510 receivers, zero baseline and short baseline tests different length. In the AF approach, only the same length using the IESSG geodetic facilities were conducted over vectors are used as desired and reference input signals. a period of two-weeks. Three kinds of Leica GPS antennas were used for two consecutive days in the short Data from the first time series, of approximately 2 minute baseline tests with a sampling rate of 10 Hz. The period, is chopped out to be compared with the second calculated time shifts are listed in Table 1. The time series with a longer period for the calculation of differences of time shifts on X and Y coordinates in each epoch’s cross-correlation coefficient. In this paper WGS84 are all within 1 second range for three types of the data used are the final coordinate time series via post- Leica antennas, but the time shift differences between Z processing. The coordinate sequence starting at epoch with X or Y could reach several seconds. Also the results tday1 of GPS time with a 2-minute interval on the first day show that the time shifts based on the day-to-day position solutions are not necessarily 4 minutes. The calculated is designated as di . The data used on the second day is day 1 maximum cross-correlation coefficients of X, Y and Z a coordinate sequence of four minutes within the interval using SR510 single frequency receivers and an AT501 of [ t +86040, t +86280]. The coefficient of cross- single frequency antennas are 0.38, 0.35 and 0.74, which day1 day1 are different from those of AT503 lightweight chokering correlation at each epoch j can be estimated by antenna, which are 0.25, 0.65, and 0.51, due to the σ )( ρ = jdx (3) Tab. 1 Time shifts for each Leica antenna in three directions (WGS84) j X Y Z σσ )()( jxjd AT501 (Sample 1) 4’06’’.3 4’06’’.9 4’07’’.8 where AT501 (Sample 2) 4’10’’.0 4’09’’.3 4’05’’.0

N AT503 4’07’’.6 4’07’’.1 4’01’’.7 j j jj σ jdx ∑ ()a a −−= xxdd ))(()( AT504 4’09’’.0 4’08’’.7 4’03’’.4 a=1 differences in signal reception and filtering.

N jj jj σ jd ∑ ()a a −−= dddd ))(()( a=1 Further studies have been made on the Z coordinates in WGS84 to evaluate the impact of misaligned data on the N jj jj efficiency of the AF approach. The approach and σ jx ∑ (a a −−= xxxx ))(()( ) a=1 algorithm used here are similar to those illustrated by Dodson et al. (2001). Figure 2 and 3 are the results from are the covariance and standard deviations of the two days’ normally (4 minutes) and exactly (4 minutes variances of two time series, respectively. and 7.8 seconds) aligned data sets using the SR510 The epoch with maximum ρ is then used as the match receiver and the AT501 antenna (Sample 1 in Table 1). The first rows in both graphs are the desired signals of the point of the two time series. Using this approach and raw AF algorithm for about a 2, 000 second period with 10 pseudorange measurement, Roberts et al. (2002) illustrate Hz sample rate, which represent the Z coordinate time the day-to-day time shifts of each PRN satellite by using series at the same location. The second rows are the time pseudorange measurements. The results reveal there are shifted reference signals of the following day’s Z differences in the time shifts for each individual satellite. coordinate time series. The third rows in Figure 2 and 3 These estimated time shifts could be employed to are the uncorrelated signal outputs, which contain mainly mitigate multipath in pseudorange measurements. A new receiver noise for this short baseline and unfiltered RINEX data file could be formed by the cleaned multipath residuals. The final rows are the correlated pseudorange measurements with the potential application components, which represent the multipath signatures. in shortening ambiguity search time. The residuals are calculated by comparing the corresponding output signals to evaluate the impact of

28 Journal of Global Positioning Systems

misaligned time series on the efficiency of the AF Residual of Receiver Noise from AF Day-to-Day Templates approach. Figure 4 and 5 are these residuals for the (From Exactly and Normally Aligned Data Sets with AT501) uncorrelated components (receiver noise plus multipath 0.015 residual) and correlated parts (multipath) of both time 0.01 series. To further investigate the impacts due to the 0.005 different antenna types, the same procedure is applied to 0 the coordinate solution from AT504 and AT503 choke -0.005 ring antennas with SR510 receivers. Noise Residual (m) -0.01

-0.015 0 2000 4000 6000 8000 10000 12000 14000 16000 18000 20000 No of Samples Fig. 5 Residual receiver noise due to misalignment (antenna AT501) Figure 6 and 7 are the residuals of multipath and receiver noise for AT504 chokering antenna. It is obvious even with a single frequency receiver and misaligned time series that if a chokering antenna is adopted for data collection, significant improvement in position solutions can be expected.

Residual of Receiver Noise from AF Day to Day Templates (From Exactly Aligned and Normally Aligned Data Sets with AT504)

0.015

0.01 Fig. 2 AF results from normally aligned Z coordinates (AT501) 0.005

0

-0.005 Noise Residual (m)

-0.01

-0.015 0 2000 4000 6000 8000 10000 12000 14000 16000 18000 20000 No of Samples Fig. 6 Residual multipath due to misalignment (antenna AT504)

Multipath Residual from AF Day to Day Templates (From Exactly and Normally Aligned Data Sets with AT 504)

0.015

0.01

0.005

0

-0.005

Multipath Residual (m) Fig. 3 AF results from exactly aligned Z coordinates (AT501) -0.01

-0.015 0 2000 4000 6000 8000 10000 12000 14000 16000 18000 20000 No of Samples Multipath Residual Calculated from AF Day-to-Day Templates (from Exactly and Normally Aligned Data Sets with AT501) Fig. 7 Residual receiver noise due to misalignment (antenna AT504)

0.015

0.01 3 Estimating Relative Displacements from Accelerations 0.005

0 The raw measurements from a triaxial accelerometer are -0.005 the discrete outputs in voltages on each axis via an Multipath Residual (m) -0.01 analogue to digital converter. These data can be further

-0.015 converted into accelerations using the zero biases and 0 2000 4000 6000 8000 10000 12000 14000 16000 18000 20000 No of Samples scale factors corresponding to each axis. To reduce the computation burden in coordinate transformations Fig. 4 Residual multipath due to misalignment (antenna AT501) between different systems, a specially designed cage is used to house a GPS antenna and accelerometer (Roberts

Roberts: Multipath Mitigation for Bridge Deformation Monitoring 29 et al. 2001). Figure 8 is the schematic of this device. It MA approach is a low-pass filter. It can be used to isolate consists of two rotary plates connected with three bolts. longer period movements from the time series of the The GPS antenna is mounted on the upper plate, which integrated velocity and displacement, which are can be orientated to north. The triaxial accelerometer is contaminated by the systematic errors of the fixed on the second plate with four screws. By rotating accelerometer. The residual between the original time the second plate one of the three axes of a triaxial series and the smoothed output from the MA constitutes accelerometer can be aligned into the same direction of a the local higher frequency variation (high-pass filter), bridge main axis (longitudinal direction). which mainly represents real structural vibration. Simulation reveals that the selection of sample number used for averaging is crucial. This requires considerable experience plus knowledge of the frequency distribution of the time series. Spectral analysis is used for the purpose of frequency identification. In the spectral analysis of the bridge data gathered from GPS and the accelerometer, the frequency of this research interest is below 2 Hz. In order to use the relative displacements calculated from accelerometer data as a reference signal to the AF approach, the accelerations are resampled from 200 Hz down to 10 Hz. In this resampling procedure, due to the low-pass nature of resampling as an MA filter, accelerometer noise or outliers with high frequency are filtered out, which can be noticed from the following data processing.

Fig. 8 A cage used to house a GPS antenna and a triaxial accelerometer 1 + s aMA t )( = ∑ ar (6) Even though the two kinds of sensors are physically qs ++ 1|| −= qr attached together, the direct comparison of GPS coordinates and accelerations using the AF approach is aMA t )( is the smoothed output for time series a at impossible due to different measurement units. The time t . a can be acceleration or filtered velocity. accelerations need to be converted into relative displacements by double integral. The accelerometer Based on the above fundamental, a Matlab script used for errors such as zero biases and scale factor errors will relative displacement calculation was developed, with accumulate into distance errors according to the equation functionalities of MA sample length selection, velocity of motion (Lawrence 1998) and relative displacement calculations and spectral analysis of output signals. += 2/1 atvtS 2 (4) Figure 9 is the comparison of original vertical In Eq. 4, v is the initial velocity; a is the acceleration accelerations sampled with 200 Hz and the resampled measurement; and S is the distance travelled in time vertical accelerations of 10 Hz. It is obvious that most period t . abnormal acceleration measurements have been filtered out through the resampling procedure. The data set Eq. 5 is used to calculate velocity from acceleration, an illustrated here is chopped from a one-hour accelerometer approach to approximate acceleration integral, data set, which was collected on 20 February 2001 on the ∆t Wilford suspension footbridge over the River Trent in tvtv )1()( +−= tata −+ ))1()(( (5) Nottingham, UK. The detailed experiment was 2 introduced by Dodson et al. (2001). To investigate the bridge dynamic responses to the pedestrians, 8 people where − tatatvtv − )1(),(),1(),( are the velocities walked repeatedly over the bridge, which caused and accelerations at time t and t −1, respectively. ∆t significant acceleration signatures. is the time interval of data sampling. Eq. 5 can also be Figure 10 and 11 are the time series of the relative applied to velocities calculated previously to output vertical velocity and displacement by using the relative displacements. aforementioned algorithm. The first graph in Figure 10 is Since the interest of the research is on the relative the integral of the resampled accelerations, which are displacements, the initial velocity can be set to zero. A illustrated by second graph in Figure 9. The second graph moving averaging (MA) filter is applied to cope with the in Figure 10 is the output of MA, which is the smoothed drift problems of the velocities and displacements in the drift caused by integral. The third graph is the calculated approximation procedure using Eq. 5. In principle, the relative velocity activated by pedestrians. Similar graphs

30 Journal of Global Positioning Systems for relative displacement are demonstrated in Figure 11. integrated sensor system of GPS receivers and triaxial The drift due to the double integral of accelerations has accelerometers could be employed. With current GPS been successfully overcome. technology, 20 Hz sampling rate can be immediately available which can be used to detect the vibration frequencies lower than 10 Hz. Under this circumstance, an integrated monitoring system of a GPS receiver and an accelerometer can provide a broader detectable frequency band except for outputting more measurements. Detailed analysis to the frequency distribution is illustrated later in the paper.

Fig. 9 Original acceleration vs. resampled ones

Fig. 12 More signatures can be identified with an integrated system

4 Acceleration Aided AF Approach for Multipath Mitigation

The relative displacements calculated from the Fig. 10 MA approach for velocity calculation accelerations by the proposed algorithm can be further used to mitigate GPS receiver noise and isolate the real bridge movements. Due to the high degree of multipath signature caused by the surroundings and receiver noise, accelerometer aided AF approach cannot be applied to GPS position solutions directly. Careful treatment is required to filter the multipath and receiver noise both at reference stations and rover stations. Meng et al. (2001) presented a hierarchical procedure to gradually reduce the impacts of multipath and receiver noise, in addition to modelling the relative tropospheric delay between reference station network. Even with such procedure, the final coordinates are still contaminated by a certain degree of the residual components aforementioned. Since

the GPS and accelerometer are independent instrument, Fig. 11 MA approach for displacement calculation the only correlated component in both systems is the Figure 12 is the result of spectral analysis on the original sensed force/movement. So, it is an ideal approach to use vertical acceleration sensed by a Kistler triaxial accelerometer aided AF technique to isolate relative accelerometer. The main vertical vibration frequencies movement of the bridge. identified by the accelerometer are 1.75 Hz, 2 Hz, 3 Hz, 5 Figure 13 is the AF results using the vertical coordinate Hz, 9 Hz, 11 Hz, and 70Hz. By the principles of spectral of two days’ as the inputs signals at the same observation analysis, the vibrations with frequencies higher than 5 Hz site. The two days’ data are exactly time shifted with the cannot be detected by a 10 Hz GPS receivers taking approach proposed in this research. In the reference time account of other error sources of GPS system alone series, there were signatures caused by casual pedestrains (Jenkins and Watts 1968). To overcome this problem, passing over the bridge but in the desired signal there either GPS receivers with a higher sampling rate or an were vibrations induced by organised crossing. Figure 14

Roberts: Multipath Mitigation for Bridge Deformation Monitoring 31 is the frequecy distribution of desired signal. The power of multipath is much higher than that of the actual bridge movement and exhibits itself as a very slow movement pattern in Figure 13. It is very difficult to identfy the excited movement without any further data treatment. From Figure 14 the dominant multipath frequencies identified by spectral anaysis are 0.0038 Hz (4 minutes 24 seconds) and 0.0075 Hz (2 minutes 13 seconds).

Fig. 16 Spectral of isolated multipath signature Figure 17 is the output using the relative multipath free GPS position solution (the third row in Figure 13 which constitutes receiver random noise and real movement) as the desired signal and the calculated relative bridge displacement from accelerometer as the reference signal in the AF algorithm. The synchronisation between the data sets of GPS and accelerometer is realised in the Fig. 13 GPS AF results using two day's position solution software whose detail is introduced by Meng (2002). The outputs of Figure 17 are the receiver noise time series (third row) and relative bridge movement (fourth row). Figure 18 illustrates the results of using untreated positioning solution from GPS directly to compare with the relative bridge movement sensed by the accelerometer. It shows that it is possible to isolate multipath from the real bridge movement but the detected real bridge movement is distorted at certain periods due to the error is too big to be adapted.

Fig. 14 Spectral distribution of desired signal Figure 15 is the spectrum of the reference signal. There is no signature of 1.75 Hz vertical movement since there was no excitation caused by organised crossing, but the same frequency signature of multipath is almost filtered out through the AF procedure (Figure 16).

Fig. 17 Acceleration aided AF result with multipath free GPS position Further analysis has been made on the frequency distributions of the input and output signals to and from acceleration aided AF system by spectral analysis. Figure 19 is the spectrum of the desired signal, which is the third row in Figure 13. It is evident that significant power reduction of multipath is realised through the AF procedure. Figure 20 is the spectrum of the reference signal of the relative displacement from resampled accelerations. It is obvious that the resample procedure Fig. 15 Spectral distribution of reference signal does not change the spectrum distribution of acceleration.

32 Journal of Global Positioning Systems

reseach is needed to identify the reason why the frequencies identified by accelerometer and within the meaurable range of 10 Hz GPS receivers have been filtered out.

Fig. 18 Acceleration aided AF result with untreated position

Fig. 21 Receiver random noise signature

Fig. 19 Spectrum of desired signal (treated GPS data)

Fig. 22 Spectrum of the integrated system

5 Summary

Fundamental of a recursive adaptive filtering (AF) approach is reviewed first in the paper. Important issues such as filter length of an AF system, convergence of AF algorithm, and time series alignment are addressed. The

focus is on the time series alignment and the consequence Fig. 20 Spectrum of reference signal (relative displacement derived of misalignment of time series to the positioning from acceleration) solutions. The exact day-to-day time shifts of the whole GPS receiver noise is the uncorrelated output via the AF satellite constellation are estimated with the GPS procedure and its spectrum is illustrated by Figure 21. It positioning solutions. A cross-correlation algorithm is provides an opportunity to further analyse the statistical used to detect the match point of two time series. The nature of GPS receiver noise. Figure 22 is spectrum of the calculated results reveal that the time shifts between X, final integrated relative displacement (Figure 17 fourth Y, and Z are different. The AF result differences by row) from the two sensors. It illustrates that the aligning two days’ time series using normal time shift (4 instrument related errors have been removed successfully minutes) and exact time shift have been compared. Two through the above data processing procedure. It needs to centimetre differences in residual noise and multipath are be pointed out that by this proposed sensor system, the evident in the positioning solution from single frequency detectable frequency is limited to the sampling frequency antenna AT501 and one centimetre for chokering antenna of GPS receivers and other signatures are left AT 504. undetectable due to the unfiltered GPS noises and the By using AF as an error suppression tool the authors weak power of these real vibration frequencies. Further demonstrate an acceleration aided AF approach to isolate

Roberts: Multipath Mitigation for Bridge Deformation Monitoring 33 real bridge movement from noise ‘polluted’ GPS GPS'97, 10th Int. Tech. Meeting of the Sat. Div. of the positioning solutions. The moving average (MA) U.S. Inst. of Navigation, 16-19 September 1997, Kansas algorithm has been applied to the original acceleration to City, USA, 1165-1172. calculate the relative displacement by double integral and Dodson A. H., Meng X., and Roberts G. W. (2001) Adaptive cope with the long period drift caused by acceleration Method for Multipath Mitigation and Its Applications for double integral. Spectral analysis is applied as a signal Structural Deflection Monitoring. International diagnosis tool to check the efficiency of the AF algorithm Symposium on Kinematic Systems in Geodesy, Geomatics in the data processing. The results illustrate that through and Navigation, 5-8 June 2001, Banff, Alberta, Canada, the AF procedure, the errors from both sensors have been 101-110. removed and real bridge movements have been isolated. Duff K., Hyzak M., and Tucker D. (1998) Real-time It is possible to achieve millimetre positioning accuracy Deformation Monitoring with GPS: Capabilities and in 3D after the removal of receiver noise and residual Limitations. SPIE 3330, Smart Structures and Materials multipath by the proposed approach. However, the 1998: Sensory Phenomena and Measurement comparison of the frequencies detected from the original Instrumentation for Smart Structures and Materials, 387- accelerometer data which are within the detectable 395. frequency band of 10 Hz GPS receivers and those from Haykin S. (2001) Adaptive Filter Theory. 4th ed. Prentice-Hall, the proposed integrated sensor system of dual frequency Upper Saddle River, New Jersey. 920p. GPS receivers and triaxial accelerometer demonstrates Ge L., Han S., and Rizos C. (2002) GPS Multipath Change that some of the detectable components were filtered out. Detection in Permanent GPS Stations. Survey Review, The possible reason for this is that these detectable 36(283): 306-322. frequencies have been removed during MA procedure Jenkins G. and Watts D. (1968) Spectral Analysis and Its which is illustrated very clear in the relative displacement Applications. Holden-Day, California. 525p. illustrated by Figure 20. Since there are no such components in the relative displacements calculated from Lawrence, A. (1998) Modern Inertial Technology: Navigation, accelerations with those from GPS measurements, only Guidance, and Control. 2nd ed. Springer, New York. 278p. the mode shape charaterised the same vibration of the Meng X., Roberts G. W., and Dodson A. H. (2001) Error first natural frequency is detected by AF technique. Analysis and Positioning Precision Improvement in Further research is needed to investigate the possible GPS/Accelerometer Structural Monitoring. Poster of IAG reasons for this and associate data processing techniques. 2001 Scientific Assembly, 2-7 September 2001, Budapest, Hungary. Meng X. (2002) Real-time Deformation Monitoring of Bridges Acknowledgements Using GPS/Accelerometers. PhD Thesis, The University of Nottingham. 239p. Leica Geosystems Ltd. (UK) sponsored the initial work in this Ogaja C., Rizos C., Wang J., and Brownjohn J. (2001) Toward area through a scholarship to support Xiaolin’s PhD research the Implementation of on-line Structural Monitoring study at the IESSG. The UK’s Engineering and Physical Using RTK-GPS and Analysis of Results Using the Sciences Research Council (EPSRC) is acknowledged for Wavelet Transform. Deformation Measurements and supplying a three-year grant to the authors to conduct research Analysis, 10th International Symposium on Deformation into using integrated GPS and accelerometers to evaluate the Measurements, 19-22 March 2001, Orange, California, integrity of structures. USA, 284-293. Rizos C. (1999) Quality Issues in Real-time GPS Positioning. References IUGG Congress, Birmingham. Roberts G. W., Dodson A. H., and Ashkenazi V. (1999) Twist Ashkenazi V., Dodson A., Moore T., and Roberts G. W. (1997) and Deflection: Monitoring Motion of Humber Bridge. Monitoring the Movements of Bridges by GPS. ION GPS World, 10(10): 24-34.

Journal of Global Positioning Systems (2002) Vol. 1, No. 1: 34-39

3D Multi-static SAR System for Terrain Imaging Based on Indirect GPS Signals

Yonghong Li, Chris Rizos School of Surveying and Spatial Information System, University of New South Wales, Sydney NSW 2052, Australia e-mail: [email protected], [email protected]; Tel: +61(2)9385-4205; Fax: +61(2)9313-7493

Eugene Donskoi, John Homer, Bijan Mojarrabi School of Information Technology and Electrical Engineering, University of Queensland, Brisbane QLD 4072, Australia e-mail: [email protected], [email protected], [email protected]

Received: 18 March 2002 / Accepted: 16 June 2002

Abstract. A 3D multi-static SAR imaging system which change detection. Examples of remote sensing utilises reflected GPS signals from objects on the Earth's applications are ocean altimetry, wind speed/direction surface is described in this paper. The principle of bistatic determination, monitoring of sea ice condition, and for radar is used to detect movement of, or changes to, the the determination of soil moisture content [1~7]. Analysis imaged object. The indirect GPS signals are processed by of indirect GPS signals has recently attracted a lot of a match filter with the aim of improving the spatial attention because of its potential civilian/military resolution of detection. The measure of spatial resolution applications. [8~13] established the models for the of this imaging system is derived, and is confirmed by extraction of sea state and wind speed from ocean MATLAB simulation. Several scenarios are considered, reflected GPS signals, and carried out some reflected for the visible satellite at a given receiver and object GPS experiments. A parallel delay mapping GPS receiver location. The scenarios for different satellites are: a) static on an aircraft was used to confirm the modelling. [14] receiver with two targets which move with the same made ocean altimetry measurements using reflected GPS speed; and b) moving receiver with one static target and signals observed from a low-altitude aircraft. one moving target. Simulation results show that the The irradiated power of GPS satellites can also be reused spatial resolution of detection depends on the relative for imaging, based only on the analysis of indirect GPS positions of the GPS satellites, the imaged objects and the signals. Generalising the bistatic radar concept, this paper GPS receiver, as well as their respective velocities. describes a multi-static synthetic aperture radar (SAR) system consisting of a constellation of visible GPS Key words: Detection, Imaging, GPS, SAR satellite transmitters, a multi-channel modified GPS receiver and multiple objects. The 'objects' may be one or more moving platforms such as a ship, or a reflective surface that is monitored for its movement. This imaging system has the following useful properties: (a) no dedicated signal transmitter is required; (b) the GPS 1 Introduction signal frequency is reused; (c) GPS operates round-the- clock and its signals cover the entire Earth's surface; (d) The Global Positioning System (GPS) is an all-weather, low power consumption; and (e) known GPS signal global, satellite-based, round-the-clock Global structure. That is, the multi-static SAR system has the Navigation Satellite System (GNSS). Measurements on potential to develop high quality, and low cost images, of the direct GPS signals have been successfully used in a localised area. navigation and positioning, while indirect or reflected signals are viewed as a nuisance. However, scattered/reflected GPS signals also can be 'reused' for remote sensing, radar target detection and (reflector)

Li: 3D Multi-static SAR System for Terrain Imaging 35

2 3D Multi-static SAR System Model frequency modulated (FM) GPS signal. The received indirect GPS signal at receiver R will be an The terrain imaging system provides visual approximately linear FM GPS signal. It has been discrimination within the image scene. A common demonstrated that the range resolution of radar is an measurement of the ability for spatial discrimination inverse ratio of the bandwidth of the signal. Hence this between objects is the spatial resolution. [15] has makes it possible for the multi-static SAR imaging described the resolution equations for a 2D configuration system to get an enough acting range and resolution in which transmitter trajectory, imaged object, as well as simultaneously, since the linear frequency modulated receiver are in the same plane. The bistatic SAR principle GPS signal has the good property of pulse compression. is traditionally based on the radar positioned on an airborne platform. The transmitter and receiver are on the same moving platform while the imaged object is static. 3 Imaging Resolution When the radar moves, the reflected signal from the same imaged object is processed in order to synthesise an In such a multi-static SAR imaging system, indirect GPS antenna with a synthetic aperture. signals that have been reflected from objects are used for In this paper, a 3D multi-static SAR system, as illustrated their detection using the bistatic radar principle. The spatial resolution is enhanced by the synthetic aperture in Figure 1, is set up as a terrain imaging system, where transmitter and receiver are located in two separate radar (SAR) technique. positions and move with different speeds. There could be As shown in Figure 1, suppose the ranges between Tri one or more imaged objects with different velocity and Oj and between Oj and R at the beginning of the vectors. The visible GPS satellites (Tr) at the receiver observation period are represented by R1ij and R 2 j position (R) act as a series of continuous signal respectively. During the period of measurement, the transmission sources. The i-th visible GPS satellite Tr i corresponding ranges are varied with respect to time t and moves with velocity V . V and V express the Tri o j R are represented by 1ij tr )( and 2 j tr )( respectively. velocity vectors of the j-th imaged object Oj and the GPS receiver R respectively. R is placed near the object Oj. All r (t) VVR ⋅−+= t)( (1a) 1ij 1 i OTrij j coordinates of Tri , Oj , and R, as well as their velocity vectors, are expressed in the Earth Centred Earth Fixed r (t) VVR )( ⋅−+= t (1b) (ECEF) coordinate system. 2 j 2 ORj j

Suppose the signal which is transmitted from the GPS satellite is:

Tri [⋅⋅= expRe)()( (ctjwAtdtS )] (2)

where d(t) is the C/A code (or P code), and wc is the carrier frequency. The received signal at Oj in complex form is: )( [ α )( ] exp{ [ −⋅⋅⋅⋅⋅−= α ttjwAFKttdtS )( ]} Oij 1ij 11 ijij c 1ij (3)

where α1ij = 1ij /)()( ctrt is the time delay of the signal

from the GPS satellite Tri to object Oj, c is the speed of light, and F1ij is the scatter coefficient. K1ij is a factor

which is associated with 1ij tr )( and F1ij .

The Doppler frequency shift caused by the movement of Tri and Oj is: Fig. 1 3D Multi-static SAR System Model 11  2  tf )( ⋅−≈ VV Rt VV ⋅−⋅+⋅− cosγ For the terrain imaging application, the bistatic radar 1ij  i OTr j 1 i OTrij j Trij  λ R   principle and the synthetic aperture radar technique are 1ij used in the signal processing. If Tri , Oj , and R move (if R VV ⋅−>> t ) (4) 1ij i OTr j with constant velocities during the observation period, the signal obtained at the Oj position will be a linear

36 Journal of Global Positioning Systems where γ is the angle between V and V , and λ is 2π 2 2π 2 Trij Tri o j k −≈ VV −−⋅ −⋅ VV (11) aij i OTr j OR j the wavelength of the GPS signal. Then R receives the λ ⋅ R1ij λ ⋅ R2 j reflected GPS signal from object O : j Since the characteristic of the correlation function near Rj )( [ 1ij αα 2 j )()( ] 22 jj AFKtttdtS ⋅⋅⋅⋅−−= the peak of the compressed wave is of interest in this discussion of resolution, let Ts >> τ . So: { Oij [ −⋅ α 2 j ()(exp tttjw )]} (5)

[SE ijnv τ ] sin)( [ aij τ ⋅⋅≈ Tkc s 2/ ] (12) where α 2 = 2 jj /)()( ctrt and Oij tw )( = 2π [ +⋅ 1ijc tff )( ]. The spatial resolution in directions x and y is: The Doppler frequency shift caused by the relative movement of O and R is: j −+− VVVV λ i j j xORxxOxTr ρ ≈ (13a) 1 d  2 j tr )(  (6) x 2 2 2ij tf )( ⋅−=  O tw )( ⋅  Ts 2π dt ij c V − OTr − VVV OR   i j + j R R where γ is the angle between V and V . 1ij 2 j R j o j R −+− VVVV The frequency at R is: λ i j j yORyyOyTr ρ ≈ (13b) y 2 2 )( ++= tftfftf )()( (7) Ts Rij 1ijc 2ij V − OTr − VVV OR i j + j Thus, the received signal at R is: R1ij R2 j

R )( [ 1ij αα 2 j )()( ] 22 jj AFKtttdtS ⋅⋅⋅⋅−−= The authors have simulated such an imaging system, with ij the transmitters, receiver, and objects moving at different  t  π )(2exp ⋅ dttfj (8) velocities.  Rij 1   ∫0 

2 As the coefficients of items with t can be ignored, 4 Simulations tS )( can be considered as an approximately linear FM Rij signal. In order to improve the bearing resolution, tS )( Assume that the observation time is from 00:00:00 24 Rij Sept. 2001 to 00:00:02 24 Sept. 2001. The coordinate of must be compressed. Because the auto-correlating receiver R is [ x , y , z ]=[-4641138, 2555708, - function of a linear FM signal exhibits a narrow pulse R R R property, the output wave will become even more narrow 3539132]m. At this time the distribution of visible GPS satellites is illustrated in Figure 2. As an example, the when tS )( is passed through a matching filter. (A Rij signals from satellites #2 and #4 are used in the object matching filter is also an optimum filter for signal imaging, and their characteristics are shown in Table 1. detection in a white noise environment.) Hence, a Suppose there are two targets (Oj , j=1,2) in initial matching filter is employed in the received signal locations [ xO , yO , zO ]=[-4643559, 2557041, - processing to improve the system resolution. The output 1 1 1 3534962]m and [ x , y , z ]=[-4643607, 2557068, - of the matching filter is: O2 O2 O2

T 3534879]m respectively. s −τ 2 * Tab. 1 Position & Velocity Parameters of Satellite #2 and #4 During Sij (τ ) = T R τ ⋅+ R )()( dttStS (9) ∫− s ij ij Observation Time 2 GPS Sat. No. #2 #4 where T is the observation time. It can be demonstrated s Position ( -25.15, (-23.57, that the normalised envelope Env [•] of Sij τ )( is: [ xTr , yTr , zTr ] -8.04, 12.33, 6 T 10 m -4.15 ) -1.90) s −τ 2 Velocity ( -4.05, ( 1.07, []SE ijnv (τ ) = T exp()aij τ ⋅⋅⋅ dttjk ∫− s 2 (VTrx, VTry, VTrz) -5.41, -3.12, 102 m/s 30.23 ) -31.83 ) 1  sin  aij Tkc s −⋅⋅= ττ )(  (10) 2  Figure 3 shows the simulation results for each scenario in Table 2. Figure 3(a)(b) and (c)(d) show that the resolution where ρ is related to the relative velocities of Oj and R. For a static receiver and moving target with speed (0,10,0)m/s, its spatial resolution is the same as that for a static target

Li: 3D Multi-static SAR System for Terrain Imaging 37 with receiver moving with velocity (0,-10,0)m/s. The indicated in Figure 3(a)(c) and (b)(d), ρ is also a function greater the relative speed, the higher the resolution. As of the position of Tri with respect to R and Oj.

Fig. 2 Distribution of Satellites During Observation Time Tab. 2 Scenarios in the Simulation

object 1 object 2 GPS Sat. (VRx, VRy, VRz) (VO1x, VO1y, VO1z) (VO2x, VO2y, VO2z) No. m/s m/s m/s ρo1x(m) ρo1y(m) ρo2x(m) ρo2y(m) 2 (0,0,0) (0,10,0) (0,10,0) 113 154 113 154 2 (0,-10,0) (0,0,0) (0,7,0) 113 152 105 142 4 (0,0,0) (0,10,0) (0,10,0) 26 79 26 79 4 (0,-10,0) (0,0,0) (0,7,0) 26 76 24 73

Fig. 3: Simulation Result

38 Journal of Global Positioning Systems

Fig. 3 Simulation Result (Continuous)

Li: 3D Multi-static SAR System for Terrain Imaging 39

5 Concluding Remarks Application, IEEE Trans. on Geosci. & Remote Sensing, 38(2), 951-964. This paper describes a 3D multi-static SAR imaging Komjathy A., et al (2000), Towards GPS Surface Reflection system using the reflected GPS signals. The bistatic radar Remote Sensing of Sea Ice Conditions, Sixth Int. Conf. on principle is used to detect movement of, or changes to, Remote Sensing for Marine & Coastal Environments, II: the imaged object. The indirect GPS signals are processed 447-456. by a match filter with the aim of improving the spatial Masters D., et al (2000), GPS Signal Scattering from Land for resolution of detection. The measure of spatial resolution Moisture Content Determination, Proceedings of IEEE of this imaging system is derived, and is confirmed via International Geoscience. & Remote Sensing Symposium, simulation studies. vol.7: 3090-3092. The simulation results show that the indirect GPS signals Garrison J.L. & S.J. Katzberg (1997), Detection of Ocean can be used for certain remote sensing applications. In Reflected GPS Signals: Theory and Experiment, IEEE such multi-static SAR imaging system the detection Southeastcon Blacksburg'97 Engineering New New based only on the reflected GPS signal can be made for Century, USA, 12-14 April, 290-294. moving objects and a moving receiver. The spatial Zavorotny Z.U., et al (2000), Extraction of Sea State and Wind resolution is a function of the mutual positions and Speed from Reflected GPS Signals: Modeling and velocities of the satellites, imaged objects, and receiver. Aircraft Measurements, IEEE Int. Geosci. & Remote The 3-D multi-static SAR model also has potential Sensing Symposium, 4: 1507-1509. benefits in sea surface imaging and target detection. Komjathy A., et al (1998), GPS Signal Scattering from Sea Surface: Comparison Between Experimental Data and Theoretical Model, Fifth Int. Conf. on Remote Sensing for References Marine & Coastal Environments, 1: 530-539. Komjathy A., et al (2000), GPS Signal Scattering from Sea Armatys M., et al (2000), Exploiting GPS as a New Surface: Wind Speed Retrieval Using Experimental Data Oceanographic Remote Sensing Tool, National Technical and Theoretical Model, Remote Sensing of Environment, Meeting of the U.S. Institute of Navigation, Anaheim, 73, 162-174. California, 26-28 January, 339-347. Lin B., et al (1998), The Relationship Between the GPS Komjathy A., et al (2001), Developments in Using GPS for Signals Reflected from Sea Surface and the Surface Oceanographic Remote Sensing: Retrieval of Ocean Winds: Modeling Results and Comparisons with Aircraft Surface Wind Speed and Wind Direction, National Measurements, J. of Geophysical Research - Oceans, Technical Meeting of the U.S. Institute of Navigation, 104(C9), 20713-20727. Long Beach, California, 22-24 January. Elfouhaily T. & C. Zuffada (2000), On Deriving Near-Surface Martin-Neira M., et al(2001), The PARIS Concept: An Wind Vector Information from GPS Ocean Reflections: Experimental Demonstration of Sea Surface Altimetry Simulation and Measurements, IEEE Int. Geosci. & using GPS Reflected Signals, IEEE Trans. on Geosci. & Remote Sensing Symposium, 7: 3081-3083. Remote Sensing, 39(1), 142-150. Lowe S.T., et al (2000), An Ocean-Altimetry Measurement Garrison J.L, et al (2002), Wind Speed Measurement Using Using Reflected GPS Signals Observed from a Low- Forward Scattered GPS Signals, IEEE Trans. on Geosci. Altitude Aircraft, IEEE Int. Geosci. & Remote Sensing & Remote Sensing, 40(1), 50-65. Symposium, 5: 2185-2187. Zavorotny V.U. & A.G. Voronovich (2000), Scattering of GPS Willis N.(1991), Bistatic Radar, Artech house Inc., Norwood. Signals from the Ocean with Wind Remote Sensing

Journal of Global Positioning Systems (2002) Vol. 1, No. 1: 40-47

Accuracy Performance of Virtual Reference Station (VRS) Networks

Günther RETSCHER Vienna University of Technology Department of Applied and Engineering Geodesy Gusshausstrasse 27-29 E128/3 A – 1040 Wien, AUSTRIA E-mail: [email protected]

Received: 24 February 2002 / Accepted: 30 May 2002

Abstract. Recent developments in differential GPS data communication link to the user (usually a radio link); (DGPS) services have concentrated mainly on the the reference station software on a PC which performs reduction of the number of permanent reference stations station monitoring, DGPS data correction model required to cover a certain area and the extension of the estimation and data archiving; interfaces and possible ranges between reference and rover stations. communication links for data transfer to the user Starting from networked DGPS stations where all stations [Landau, 2000]. For integrity monitoring, a reference are linked to a central control station for data correction station usually consists of 2 independent GPS receivers to and modeling, the most advanced technique nowadays is guarantee against system failure. The user receives either based on the virtual reference station (VRS) network DGPS corrections for code positioning or real-time concept. In this case, observation data for a non-existing kinematic (RTK) GPS data for carrier phase positioning “virtual” station are generated at the control center and in RTCM (Radio Technical Commission for Maritime transmitted to the rover. This leads to a significant Services) format. As the observation errors and biases are improvement in positioning accuracy over longer not modelled in the network, the error budget shows a distances compared to conventional DGPS networks. distance dependent growth as the user-to-station This paper summarizes the various DGPS architectures separation increases [Retscher and Moser, 2001]. and the corresponding accuracy. This is followed by a description of the models and algorithms used for the VRS station concept. Practical examples of correction data services in Europe are given to highlight the achievable positioning accuracy. The results of an analysis of test data in a virtual reference station network in southern Germany show that always a horizontal positioning accuracy in the order of ± 5 cm can be achieved for baselines with a length up to 35 km.

Key words: GPS, DGPS, VRS, RTK Fig. 1 Single reference station design

1.2 Multiple station concept

In the second concept, multiple reference stations are 1 DGPS NETWORK ARCHITECTURES connected to a central control station using a data communication link (wireless radio link or cable connection via local area network (LAN) or Internet). 1.1 Single reference station concept Additional equipment at the reference station includes a modem for data transfer and modification of the station Figure 1 shows the architecture for the single reference software package. The data transfer protocol employed station concept. In this case, a reference station in a between each reference station and the control center is DGPS network consists of the following main usually RSIM (Reference Station Integrity Monitor components: a GPS antenna/receiver assembly; a wireless messages). Further information about the data format

Retscher: Accuracy Performance of VRS Networks 41 standards may be found in [Moser, 2001]. On the control concept (Figure 2), only due to the software modification station, software is used to monitor several reference for data modeling in the control center a networked stations simultaneously. solution is obtained. The software modification includes new models for correction data estimation for modeling of the major error sources, i.e., the satellite orbit errors, ionospheric refraction as well as satellite and receiver clock errors. For the modeling observation data from at least three multiple reference stations are required. Then so-called area correction parameters [Wanninger, 1999] can be deduced for each triangle of three reference stations in a network. Other advantages of a networked DGPS station network include the possibility for detecting station malfunction or failure. Figure 3 compares the positioning accuracies

which are achieved over the SATVB reference station Fig. 2 Multiple reference station design with control center network in Burgenland, Austria [Retscher and Chao, 2000] for a standard DGPS network on the left (Figure 3 (a)) and the networked solution on the right (Figure 3 1.3 Networked DGPS system concept (b)). From the comparison it can be seen that due to the networked solution an accuracy of better than ± 2 cm can Due to the use of networked DGPS system concepts, a be achieved for the whole area of Burgenland using only reduction in initialization time for carrier phase 4 reference stations where the station separation ranges positioning (i.e., time required for resolving the carrier between 40 to 50 km. In the standard DGPS network the phase ambiguities) and accuracy improvement for longer position accuracy degrades as the user-to-station ranges is achieved. Additionally the reliability of the separation increases and reaches values of ± 5 cm at a position solutions are improved allowing a larger distance of 50 km from the nearest reference station. reference-to-user separation. Thereby the system architecture is similar to the multiple reference station

(a) Standard DGPS solution for four independent reference stations (b) Networked DGPS solution with an additional central control station

Fig. 3 SATVB network in Burgenland, Austria with 4 CORGS (Contiuous Operating Reference Geodetic Stations) [after Titz, 1999]

42 Journal of Global Positioning Systems

the initialization time. 1.4 Virtual reference station network concept The system architecture is shown in Figure 4. To create the virtual reference station data for a certain RTK GPS Currently the most advanced approach for increased rover station, the user receiver’s approximate location spatial separation of permanent stations and error accurate to about 100 m is transmitted to the network modeling is the so-called virtual reference station (VRS) control center. As a result, a bi-directional communication network concept. The concept was firstly introduced in a link between the user and the control center is required. part of the German reference station network SAPOS The communication is usually performed using cellular [Landau, 2000; Trimble, 2001]. The name of this phones (Global System for Mobile Communication approach results from the fact that observations for a GSM, in future Universal Mobile Telecommunication “virtual” non-existing station are created from the real Service UMTS). The observations for a given location are observation of a multiple reference station network. This estimated in the control center using real-time correction allows to eliminate or reduce systematic errors in models and then transferred in the RTCM format to the reference station data resulting in an increase of distance rover station. On the rover side, standard RTK GPS separation to the reference station for RTK positioning algorithms are employed to obtain the position fix. while increasing the reliability of the system and reducing

Fig. 4 System architecture of the virtual reference station concept [after Trimble, 2001] the rover leaves a VRC cell, then it is assigned to a new cell. The main advantage using this concept is that there 1.5 Virtual reference cell concept is no limitation in the number of users and no bi- directional communication link between the rover and the Another possibility for networked DGPS solutions is the control center is required. The positioning accuracy, virtual reference cell (VRC) concept. Here the correction however, is lower than the accuracy which can be models are not estimated for a specific user on request as achieved using the VRS approach. The VRC concept is in VRS concept, but the models are estimated for a given therefore mostly employed in WADGPS (Wide Area gridded DGPS service area. The rover receiver is DGPS) networks, such as the services provided by Thales assigned to a cell and there is no need for the virtual and Fugro. Further information about this services may station to follow the movement of the rover station. When

Retscher: Accuracy Performance of VRS Networks 43 be found in e.g. [Moser, 2001], [Retscher and Moser, reference stations. In the dynamic orbit determination 2001]. algorithm also ionospheric correction parameters can be estimated in an integrated Kalman filter approach. Thereby the TEC (Total Electron Content) is estimated 2 ERROR BUDGET AND MODELING using modified standard single layer models (e.g. a modified Klobuchar model) [Klobuchar, 1986; The main error sources and the models employed in Kleusberg, 1998]. The clock errors are depending on the WADGPS networks have been discussed in detail in other two error sources. They are estimated in an iterative [Retscher and Chao 2000]. The models can be adapted to process and due to the improvement of the estimates of suit the requirements for correction data estimation in the other error sources their impact is reduced. Site- networked DGPS services or in the virtual reference specific errors (e.g. multipath) can not be taken into station network concept. As usual, the main error sources account as it has to be assumed that the reference stations that have to be dealt with are the satellite orbit errors, are situated at ideal locations and at the user site these ionospheric refraction as well as satellite and receiver error sources are also minimized. clock errors. For the error budget see [Kaplan, 1996] and for a detailed discussion of the error modeling see [Ashkenazi et al., 1997; Retscher and Moser, 2001; 3 RESULTS OF TEST IN A VIRTUAL REFERENCE Whitehead et al., 1998]. In general, the approaches can be STATION NETWORK classified into the following three types: The performance and accuracy achievements of the • Estimation of range corrections: In this case, networked virtual reference station concept were analysed and solutions are used to estimate the range corrections presented in [Retscher and Moser, 2001]. The main from weighted observations of a multiple reference results are summarized here. The VRS test network in station network. The main disadvantage is that the southern Germany of the company Trimble (Terrasat) is positioning accuracy gets worse depending on the shown in Figure 5. For the analysis presented here, the distance between the user and the central point of the observation data for three stations (Virtuell 1 to multiple reference station triangle. Virtuell 3, see Figure 6) in the VRS network was 1 • Estimation of position corrections: Position dependent downloaded from the website of the company Terrasat . algorithms estimate the position correction from the In total, 112 measurement epochs have been processed. weighted average of the rover positions derived from To analysis the distance dependance of the result, the each reference station. The accuracy degrades in a station Höhenkirchen was chosen as reference station in similar way to the method using the range correction all tests. In the first analysis, the accuracy of the solution approach. was investigated, followed by an analysis of system performance and the overall precision of the result. • Estimation of corrections for the cell area: In the third approach, the range error is divided into different components that are estimated in a cell area independently from the baselines between the reference and rover stations. The user receives DGPS corrections that are separated into several components and must combine them to determine a position solution. The calculation of the corrections is an iterative process. This approach is most commonly applied and has many advantages compared to the first two algorithms named above. The disadvantage, however, is that the corrections have to be applied at the rover and a modification of standard DGPS algorithms would be necessary. This problem has been solved using the VRS approach. For real-time applications, these error models have to be predicted ahead. An approach for a dynamic orbit determination has been introduced in [Retscher and Chao, Fig. 5 VRS test network of the company Trimble (Terrasat) in Bayern, 2000]. It can be summarized that then the satellite orbits Southern Germany (Map not to scale) can be obtained with an accuracy of better than 10 m r.m.s. using dual frequency raw pseudorange data from only three reference stations. The accuracy is further 1 improved using data from more than three multiple Website for Download of VRS observation data: www.virtualrtk.com

44 Journal of Global Positioning Systems

Fig. 6 Virtual reference station locations (Virtuell 1 to Virtuell 3)

3.1 Accuracy of the solution for the VRS stations

Table 1 shows the standard deviations of the processing results for the VRS stations (Virtuell 1 to Virtuell 3). The observations have been treated as kinematic observations and positions were processed independently for each measurement epoch. Figure 7 shows a classification of the standard deviations of the horizontal coordinates using a class interval of 10 cm. As expected, it can be seen from Table 1 that the standard deviation increases over larger distances from the reference station (i.e., station Höhenkirchen). On the other hand, surprisingly, the standard deviations of station Virtuell 2 and Virtuell 3 Fig. 7 Standard deviations of horizontal coordinates of the virtual have nearly the same value, although the station reference stations (classification with intervals of 10 cm) Virtuell 3 is 18 km further away from the reference station Höhenkirchen than Virtuell 2. The reason for this phenomenon may be the fact that the station is located 3.2 Overall precision of the result for the VRS stations outside the triangle of the three multiple reference stations (see Figure 6) which are used to estimate the The overall precision of the result was obtained by correction parameters in VRS reference station network. comparing the solution for the VRS stations with the true values of the coordinates used to download the observation data. Figure 8 shows a classification of the Tab. 1 Standard deviations of the kinematic solution for the VRS stations Virtuell 1 to Virtuell 3 deviations of the horizontal components X and Y for the stations Virtuell 2 and Virtuell 3 where again a class Standard deviations in [cm] interval of 10 cm is used. As expected, the deviations for Point No. horizontal coord. Height station Virtuell 1 are very small and are not displayed in Virtuell 1 +/- 2.0 +/- 4.3 Figure 8. The deviations follow a Gaussian distribution which proves that no systematic errors occur in the data Virtuell 2 +/- 34.4 +/- 65.2 sets. Virtuell 3 +/- 37.1 +/- 68.8

Retscher: Accuracy Performance of VRS Networks 45

(a)

(b)

Fig. 8 Deviations of the horizontal component for the VRS stations Virtuell 2 (a) and Virtuell 3 (b) (classification with intervals of 10 cm) The following major results can be summarized: station is given by the standard deviation of their differences to the true values. • The baseline accuracy of RTK GPS measurements is usually described by a constant and a distance • For a comparison of all results the standard deviations dependent error, e.g. 5-20 mm ± 1-2 ppm. For a of the differences to the true values of the VRS stations baseline with a length of 10 km we would therefore get are summarized in Table 2. To achieve comparable an error of ± 40 mm in the worst case. In the analysed values at a probability level of 99%, the standard concept of the VRS network, the maximum baseline deviations of the measurements of 112 epochs have to length is always very short as observations of a non- be multiplied by a quantile of 1.211 which is obtained existing “virtual” reference station are sent to the rover from a student probability distribution. The results station which is located nearby the rover. The baseline show reasonable values for the standard deviations of length is given by the square root of the square sum of the X and Y coordinates for distances up to 30 km from the coordinate differences between the VRS and the the network center point. In addition, the standard rover station. In our investigation the maximum deviations are also compared to values published by the distance encountered was less than 1.05 m. Therefore company Trimble (Terrasat) for the station Neufahrn. the precision of the position solutions for the rover They have been obtained from a continuous RTK stations can be equated with the precision of the observation over a period of several hours. As can be corresponding VRS station. The precision of the VRS seen from Table 2 similar results are obtained for the horizontal component, the standard deviations of the

46 Journal of Global Positioning Systems

height component, however, are much smaller. The the standard deviation as the length of the observation reason for this may be the fact that a larger number of period was about 90 hours. RTK results are available which are used to calculate Tab. 2 Comparison of the standard deviations of four VRS stations Standard deviations in [cm] of the differences Distance to the Point No. at a probability level of 99% network center point X Y H 8km Virtuell 2 <2.2 <2.9 <13.1 13km Neufahrn <2.6 <2.1 < 4.9 24km Virtuell 1 <1.6 <0.4 <10.1 27km Virtuell 3 <2.3 <2.9 <13.0

• The achievable precision for the horizontal component (German). http://info.uibk.ac.at/c/c8/c802/obg99/ahrer. of the solutions are always within ± 5 cm, even for html (Last visited: May, 2001) baseline length up to 35 km. Also for the height good Ashkenazi, V.; Chao, C. H. J.; Chen, W.; Hill, C. J. & Moore, results can be obtained where as usual the standard T.: A New High Precision Wide Area DGPS System. The deviations are larger by a factor of 1.5 to 2 compared to Journal of Navigation, Vol. 50, No. 1, January 1997, pg. the horizontal component. Therefore our tests could 109-119. prove that a high precision increase can be achieved Chao, C. H. J.: An Integrated Algorithm for Effective Orbit due to the employment of network station concepts. Determination. The Geomatics Journal of Hong Kong, Vol. 1, No. 1, July, 1997, pg. 53-62. Chao, C. H. J. & Ding, X.: Single Frequency Ionospheric 4 CONCLUSIONS Modelling for Wide Area Differential GPS. in: Papers presented at the XXI International Congress FIG’98, Using the VRS station concept, similar accuracies can be Brighton, UK, July 19-26, 1998. achieved in distances of up to 35 km from the nearest Kaplan E. D.: Understanding GPS – Principles and reference station as for short baselines in the single Applications. Artech House Publishers, Boston, London, reference station concept. Therefore the distances 1996. between the reference stations in a LADGPS network can be enlarged to 70 to 80 km which would result in large Kleusberg, A.: Atmospheric Models for GPS. in: Teunissen, P. cost savings for the establishment and a maintenance of J. G. & Kleusberg, A. (Eds.): Chapter 15 of the book GPS for Geodesy, Springer Verlag, Berlin, Heiddelberg, New the permanent DGPS network. A further advantage of the York, 1998, pg. 599-624. VRS concept is that in the rover receiver standard RTK processing algorithms are employed and no modification Klobuchar, J. A.: Design and Characteristics of the GPS of the receiver hardware or software is required. The Ionospheric Time Delay Algorithm for Single Frequency communication link is performed using common mobile Users. in: Proceedings of the PLANS-86 conference, Las Vegas, 1986, pg. 280-286. phone data links. Due to high density of mobile transmitters in Europe nearly a full coverage of most Landau, H.: Zur Qualitätssicherung in GPS areas is guaranteed. For a global use of the data Referenzstationsnetzen, Paper presented at the XXIII communication, however, a modification of the Course of Engineering Surveying, Munich, March 13-17, commonly used RTCM data protocol is still required. It Verlag Konrad Wittwer, Stuttgart, pp. 277-290 (German). can be expected that this problem will be solved soon. Moser, R.: Untersuchung und Vergleich von Local Area und New networks in Austria, e.g. a new permanent DGPS Wide Area DGPS Diensten. Diploma thesis, Department network for Vienna which will be established by the of Applied and Engineering Geodesy, Vienna University of power supply company Wienstrom, will employ the most Technology, 2001 (German). advanced VRS station concept. Retscher, G.: RTK-GPS Positioning and Navigation in Marine Geodesy. The Geomatics Journal of Hong Kong, No. 2, 1999, pp. 39-48. REFERENCES Retscher, G. &. Chao, C. H. J.: Precise Real-time Positioning in WADGPS Networks. GPS Solutions, Vol. 4, No. 2, Fall Ahrer, H. & Auzinger, T.: Der österreichische DARC-DGPS 2000, pp. 68-75. Dienst. in: Papers presented at the „10. Internationale geodätische Woche Obergurgl“, 21-27 February, 1999,

Retscher: Accuracy Performance of VRS Networks 47

Retscher, G. & Moser, R.: Untersuchung und Vergleich von Engineering, April 20-22, 1998, Eisenstadt, Austria, pg. Local Area und Wide Area DGPS Diensten. Allgemeine 564-569. Vermessungsnachrichten, No. 10, October 2001 (German). Trimble (Terrasat): Introducing the Concept of Virtual Teunissen, P. J. G.: GPS Carrier Phase Ambiguity Fixing Reference Stations into Real-Time Positioning. Technical Concepts. in: Teunissen, P. J. G. & Kleusberg, A. (Eds.): information, 2001. http://www.terrasat.com/applications/ Chapter 15 of the book GPS for Geodesy, Springer Verlag, refvirtual.htm (Last visited: January, 2002) Berlin, Heiddelberg, New York, 1998, pg. 319-388. Whitehead, M. L.; Penno, G.; Feller, W. J.; Messinger, I. C.; Titz, H. & Weber, R.: SATVB – A multipurpose Bertiger, W. I.; Muellerschoen, R. J.; Iijima, B.A. & GPS/GLONASS reference station network in Piesinger, G.: Satloc Real-Time Wide Area Differential Burgenland/Austria. in: Papers presented at the GPS System. GPS Solutions, Volume 2, Number 2, 1998, Symposium on Geodesy for Geotechnical and Structural pp. 46-63.

Journal of Global Positioning Systems (2002) Vol. 1, No. 1: 48-56

Pseudolite Applications in Positioning and Navigation: Progress and Problems

J. Wang School of Surveying and Spatial Information Systems, University of New South Wales, Sydney, NSW 2052, Australia e-mail: [email protected]; Tel: +61(2)9385 4203; Fax: +61(2)9313 7493

Received: 4 May 2002 / Accepted: 18 July 2002

Abstract. Global navigation satellite systems have been such as in urban canyons and deep open-cut mines, the revolutionising surveying, geodesy, navigation and other number and geometry of visible satellites may not be position/location sensitive disciplines. However, there are sufficient to reliably carry out the positioning operations. two intrinsic shortcomings in such satellite-based In the worst situations, such as underground or inside positioning systems: signal attenuation and dependence buildings, the satellite signals are completely lost. Such on the geometric distribution of the satellites. problems with space-borne satellite positioning systems Consequently, the system performance can decrease can be addressed by additional ranging signals significantly under some harsh observing conditions. To transmitted from ground-based "pseudo-satellites" tackle this problem, some new concepts of positioning (pseudolites). with the use of pseudo-satellites have been developed and The concept of the pseudolite was proposed in the tested. Pseudo-satellites, also called pseudolites, are 1970’s, even before the launch of the GPS satellites. In ground-based transmitters that can be easily installed fact, pseudolites were originally designed to test the wherever they are needed. They therefore offer great initial GPS user equipment (Harrington & Dolloff, 1976). flexibility in positioning and navigation applications. During the past decade, new pseudolite concepts and Although some initial experimental results are hardware have been developed for a variety of encouraging, there are still some challenging issues that positioning and navigation applications. Pseudolites can need to be addressed. This paper reviews the historical be used as an augmentation tool for space-borne satellite pseudolite hardware developments and recent progress in positioning systems. This augmentation can improve the pseudolite-based positioning, and discusses the current system performance because the availability and technical issues. geometry of positioning solutions are significantly strengthened. Further more, a pseudolite-only positioning Key words: GPS, Pseudolites, Indoor Positioning, system is possible, which can replace the space-borne Navigation satellite constellation where the use of satellite signals is not feasible, such as underground and indoors.

This paper presents an overview of the pseudolite hardware developments and recent progress in pseudolite positioning applications, and discusses the current 1 Introduction challenging issues, such as pseudolite and receiver hardware development, pseudolite synchronization, Global navigation satellite systems, such as GPS and multipath effects and modelling errors. Glonass, have been playing an increasingly important role in surveying, geodesy and navigation, in which positioning is a major component. It is well known that 2 Hardware Developments for such space-borne satellite positioning systems the accuracy, availability and reliability of the positioning During the early days of GPS development, the test results is very dependent on both the number and facility, the Inverted Range, was established. In this test geometric distribution of satellites being tracked. range at Yuma Proving Ground (USA) four ground However, under some harsh observing environments, transmitters provided the simulated GPS satellite signals

Wang: Pseudolite Applications in Positioning and Navigation 49 for testing GPS receivers. These ground transmitters (Söderholm et al., 2001). These pseudolites transmit (GTs) were so-called pseudolites (Harrington & Dolloff, GPS-like ranging signals. For this reason, they are called 1976). These first pseudolites were designed to transmit GPS pseudolites (Elrod & Van Dierendonck, 1996). They GPS L1 signals, although the navigation message for can be programmed or preset to broadcast any of the gold these pseudolites was different from that for the GPS codes of GPS (i.e., PRN codes from 1 to 37) on GPS L1 satellites. In fact, the pseudolites just transmitted the at the frequency of 1575.42Mhz. Some types of GPS fixed coordinates for the pseudolite locations (while the signal simulators, such as the Stanford Telecom Model other portions of the navigation message for the satellites 7201 wideband Signal generator (Holden & Morley, are not applicable in the case of pseudolites). 1997) and GSS simulators (Weiser, 1998), can be configured to transmit a GPS-like C/A code signal on L1. The pseudolite concept developed during the GPS Thus, these GPS signal generators/simulators can development stage is being used again during the current essentially be used as a pseudolite. GPS modernization programme (ITT, 2002). A new Inverted GPS Range (IGR) will be developed at In principle, pseudolites can transmit their ranging signals Holloman Air Force Base (USA) to support validation of on different frequencies, just as the GLONASS satellites new military and civilian signals planned for the do. Australia’s CSIRO Telecommunications and modernised GPS constellations. This new GPS user test Industrial Physics is currently developing a high precision facility will be open to both military users and location system (PLS) which uses the ISM band manufacturers of civilian receivers. frequencies (http://www.tip.csiro.au/ICT/PrecisionLocator /index.htm). Zimmerman et al. (2000) proposed a Similar to the GPS, a new global navigation satellite pseudolite design that uses up to five frequencies (two in system GALILEO under development by the European the 900MHz ISM band, two in the 2.4GHz ISM band, and Union will also use the pseudolite concept to test and the GPS L1 frequency). An advantage of such multi- validate frequency allocation and user equipment. A frequency pseudolite systems is that the integer carrier GALILEO pseudolite is currently being developed by the phase ambiguities can be resolved instantaneously, due to Institute of Geodesy and Navigation (IfEN) at the redundant measurements and the extra wide-lane University FAF Munich (Hein, 2002). observables that can be constructed from these The use of pseudolites in positioning and navigation was frequencies. first discussed by Beser & Parkinson (1982) and Klein & New pseudolite hardware designs have been proposed Parkinson (1984). In the mid 1980s, the Radio Technical during the past few years. The latest hardware designs are Commission for Maritime Services (RTCM) defined a closely connected to new applications. Some examples pseudolite which can receive GPS satellite signals, are: compute pseudorange and range-rate corrections, and transmit the correction information at 50 bits per-second • In order to use the pseudolite signals in single point on an L-band frequency. In addition, the transmitted positioning, it is necessary to synchronise the signal should be GPS like and the signal is designed to pseudolite ranging signals to the GPS signals. This prevent interference to GPS and other equipment. The kind of pseudolite is called a Synchrolite (Cobb, 1997). RTCM committee SC-104 ('Recommended Standards for Differential Navstar GPS Service') designated the Type 8 • To implement a Mars pseudolite array navigation Message for the pseudolite almanac, containing the system as proposed in LeMaster & Rock (1999), the location, code and health information of pseudolites pseudolites have been designed to be capable of both (Kalafus et al., 1986). However, at that time, the receiving and transmitting ranging signals at GPS development of a prototype pseudolite of the type defined L1/L2 or other frequencies. This type of pseudolite can by the RTCM was costly, with the indicative prices ‘exchange’ signals, self-determining the geometry for a ranging from US$100K-200K (Parkinson & Fitzgibbon, pseudolite array. These pseudolites are referred to as 1986). Transceivers. Stone et. al. (1999) have reviewed transceiver applications. In the early 1990s, researchers at Stanford University developed a low cost GPS L1 C/A code pseudolite for • A pseudolite can be installed on stratospheric use in a CAT III automatic landing system (Cohen et al., platforms, as shown in Figure 1, to broadcast both 1993). During the past decade, commercial pseudolite ranging signals and differential corrections for GPS, hardware products have become available on the market. GLONASS and GALILEO systems (Dovis et al., In the mid 1990s, the first commercial pseudolite product 2000). Such a pseudolite design is called Stratolite. was manufactured by the IntegriNautics company Currently the majority of the pseudolites transmit GPS- (www.integriNautics.com). In 2001, another like signals at the L1 frequency (1575.42MHz) and manufacturer, Navicom, launched a new pseudolite possibly on L2 (1227.6MHz). With this configuration, product called NGS1T (http://www.navicom.co.kr). standard GPS receivers can be used to track pseudolite Another pseudolite product for indoor tracking and signals with the modification of the firmware. Currently, navigation services is under development in Finland

50 Journal of Global Positioning Systems it has been identified that NovAtel Millennium and problem. This approach is based on a signal processing Canadian Marconi Corp. Allstar GPS receivers can be technique which does not require receiver hardware used to track pseudolite signals. In addition, some GPS modification. A theoretical analysis has shown that a receiver development kits, which include receiver combination of code and phase can deal with the near- firmware source code, can be modified for pseudolite fare problem (Progri & Michalson, 2001). It is also applications. For example, the Mitel (now Zarlink) GPS reported that the special pseudolite transmitter antennas Architect 12 Channel Development Kit has been used for with appropriate radiation patterns can address the near- this purpose (e.g., LeMaster & Rock, 1999; Stone & far problem (Söderholm et al., 2001). Powell, 1998; Wawrzyniak et al., 2001)

3 Pseudolite Positioning and Navigation Applications

As ground-based radio signal transmitters, pseudolites have been used to augment the GPS constellation, to form an independent system for positioning and navigation applications, and to integrate with other sensors. During the past decade, with a variety of pseudolite hardware designs, the investigations into pseudolites have intensified across a wide range of applications.

Fig. 1 Differential satellite positioning with a stratolite 3.1 GPS Augmentation with Pseudolites (http://www.helinet.polito.it) Although pseudolites transmit ranging signals similar to The applications of pseudolites in augmenting GPS GPS satellites, the pseudolite signals can be much satellite constellation had been exploited even at the GPS stronger than GPS signals. Therefore there is a potential development phase. In fact, the first pseudolites deployed interference with the satellite signals due to the pseudolite at the Inverted Range were also used to test the transmitter(s) being very close to the receiving antenna differential GPS (DGPS) concept (Beser & Parkinson, compared to the GPS satellites. However, if the 1982). The role of pseudolites in differential GPS transmitters are too far from the receiver antenna, the applications was discussed in, for example, Kalafus et al. pseudolite signals will be too weak to be tracked. This is (1986), Parkinson & Fitzgibbon (1986), Stansel (1986), referred to as the 'near-far' problem, which is caused by and Van Dierendonck et al. (1989). In DGPS the higher dynamic range of the signal strength a user applications, a pseudolite can be used to provide not only receiver will experience when the receiver is in motion an additional ranging signal, but also a differential data within the proximity of pseudolite signal transmitters link. Holden & Morly (1997) have reported the (e.g., Cobb, 1997). development and test of a pseudolite-augmented DGPS system GuideNetTM which is built around the Stanford Klein & Parkinson (1984) have proposed three potential Telcom Model 7201 Wideband Signal Generator solutions for the near-far problem: (1) to pulse the (pseudolite). pseudolite signals at fixed cycle rates; (2) to transmit the signals at a frequency offset from GPS L1, but within the same frequency band as GPS; (3) to use different codes that have a longer sequence than the existing GPS codes. Galijan & Lucha (1993) proposed a GLONASS pseudolite concept, which is similar to solution (2). The major advantage is that the GLONASS pseudolites will have a larger near/far ratio, approximately 20 times that of C/A code GPS pseudolites. However, there are potential problems with the receiver designs, one of which is inter-channel biases varying with the antenna temperature. In addition, this may also lead to complicated modelling and ambiguity resolution procedures (Wang, 2000; Wang et al., 2001a). Because Fig. 2 A navigation and positioning service using stratolites (Tsujii et solutions (2) and (3) require modifications of the GPS al., 2001) receivers, solution (1) is the preferred choice for general applications. A concept for a navigation and positioning service using pseudolites on Japan’s airship-based stratospheric Recently, Madhani et al. (2001) proposed a successive platforms (SPF) has been proposed (Tsujii et al., 2001). interference cancellation approach to mitigate the near-far As shown in Figure 2, the airships will be deployed at an

Wang: Pseudolite Applications in Positioning and Navigation 51 altitude of about 20km, and thus the separation between pseudolite measurements, the success rate of ambiguity pseudolites (stratolites) and users varies from 20 to 70km. resolution and the reliability of positioning can be The near-far problem will not be as serious a problem as improved (Verhagen, 2001; 2002). in the cases of ground-based pseudolite applications. In this concept, pseudolite signals are considered as extra 3.2 Pseudolite-only Positioning satellite signals in navigation and positioning solutions, at both code and carrier phase levels. In principle pseudolites can replace the satellite GPS augmentation signals can be transmitted from constellation for positioning and navigation wherever airborne or space-borne platforms, such as airplanes and satellite signals are unavailable, such as indoors, spacecraft. Airborne pseudolites (APL) have been tested underground carparks, long tunnels, and even on other for military applications (Tuohino et al., 2000). Just like planets. Actually, the very first pseudolite application was GPS, such APL research development, driven by the pseudolite-only positioning (Harrington & Dolloff, military purposes, may also benefit the civil pseudolite 1976). A pseudolite-only positioning and navigation applications. With high orbits, GPS has been widely used concept has been proposed and tested for indoor in the navigation and attitude determination for low earth positioning (e.g., Zimmerman, 1996; Kee et al., 2000). orbit (LEO) spacecraft, such as the International Space The basic principle behind such an indoor positioning Station (ISS). However, for some spacecraft approaching concept is still the ‘double-differencing’ procedure as the ISS for docking and/or other operations in the vicinity used in precise GPS relative positioning. of the ISS, there may not be sufficient GPS signals for a As with satellite-based positioning systems, the reliability navigation solution, as the ISS is a huge structure, which of a pseudolite-based positioning system is very can block GPS signals. An investigation is underway as dependent on the strength of receiver-pseudolite to the feasibility of installing pseudolites on the ISS geometry in the system. A simulation has been carried structure, transmitting ranging signals to the approaching out to analyse the geometry strength for an indoor space vehicles (Wawrzyniak et al., 2001). In other application scenario, in which five pseudolite transmitter developments, pseudolites have been considered for antennas were installed on the ceiling (10 metres above onboard orbit and attitude control of geo-stationary the floor). Figure 3 shows the RDOP values for a rover satellites (Altmayer et al., 1998) and Spacecraft moving around on the floor of the room (Wang et al., Formation Flying (Corazzini & How, 1999). 2001c). In this scenario, the RDOP varies from 1.2 to 3.8, During the last decade the most notable pseudolite suggesting a reasonably good positioning geometry. application has been in aviation for precision approach and landing (e.g., Brown, 1992; Cohen et al., 1993; Hein et al., 1997; Bartone & Kiran, 2001). In these applications, pseudolite measurements can provide an extra check on the integrity of navigation solutions (Pervan et al., 1994). In addition, rapid change of the geometry between the pseudolites and the users can speed up the carrier phase integer ambiguity resolution, which is a key prerequisite for precise navigation operations. A motion-based ambiguity resolution approach has been developed and tested by Cohen et al. (1993).

It has been established that the pseudolite ranging signals Fig. 3 Indoor positioning scenario: 5 pseudolites 10m above the floor can contribute to the satellite positioning systems by enhancing the geometric strength, improving the Simulation is a useful tool for pseudolite-based availability, integrity, and reliability, and increasing the positioning design studies. A software simulation tool has accuracy of the positioning solutions, especially in the been developed to predict achievable accuracies from an height component (e.g., Morley & Lachappelle, 1998; array of six pseudolites within a tunnel (Calijan, 1996). Stone & Powell, 1998; Dai et al., 2001a; Wang et al., Such a simulation study has shown that deployment of 2001b). six pseudolites with a good geometry can potentially provide 1-5 cm horizontal positioning accuracies within a It has been demonstrated that the pseudolite carrier phase tunnel of 150m in length. These positioning accuracies measurements are of high precision, even at a very low are sufficient for vehicle tracking and control in future elevation angle (e.g., Wang et al., 2001b). Precise Automated Highway Systems (AHS). pseudolite carrier phase measurements are under study for a wide range of applications, such as machine control A novel approach has been suggested for the use of at mining sites (Stone et. al., 1999), and deformation pseudolites in Mars explorations (LeMaster & Rock, monitoring applications (Dai et al., 2001b; Barnes et al., 1999). Figure 4 shows a Mars pseudolite array, designed 2002; Meng et al., 2002). With the introduction of to provide centimetre-level location and attitude

52 Journal of Global Positioning Systems information to robotic rovers. This high accuracy receiver and reference pseudolite locations should be navigation capability is also a key technical requirement precisely pre-determined. With the known coordinates for for future astronaut/robot exploration teams to Mars. the receivers and the reference pseudolite, the coordinates for the user pseudolite can be determined. Extensive experiments have been conducted to evaluate the performance under various operating environments (e.g., O’Keefe et al., 1999; Tsujii et al., 2001; Dai et al., 2002; Barnes et al., 2002). As with satellite-based positioning systems, the reliability of an inverted pseudolite positioning system is very dependent on the strength of geometry of the receivers and pseudolites used in the system. For an inverted pseudolite positioning system, a poor geometry will be one in which all the receivers and pseudolites lie Fig. 4 Mars Pseudolite Array (LeMaster & Rock, 1999) approximately in the same plane (Pachter & Mckay, Just like other pseudolite-based positioning systems, the 2000). Such poor geometry will amplify the errors in the locations of the pseudolites need to be precisely positioning solutions. Also, ‘unfavorable’ geometry may determined. However, this will be a difficult task when occur, for example, when the transmitter antenna stays placing the pseudolites on another planet like Mars. To directly over the centre of a planar four-receiver square. address such a challenge, a new pseudolite positioning In this situation the design matrix of the measurement system, a Self-Calibrating Pseudolite Array (SCPA), has equations will become singular, and thus there is no been proposed by the Aerospace Robotics Laboratory at unique positioning solution. Such situations can be Stanford University (LeMaster, 2002). As mentioned identified through a full simulation for all possible earlier, the pseudolites used in such a system are actually trajectories and excluded from positioning operations. transceivers. The transceivers can transmit and receive ranging signals to determine the relative locations of all 3.3 Integration of GPS, Pseudolites and INS the pseudolites in the array. As shown in Figure 5 (LeMaster, 2002), a transceiver consists of a transmitter Given the flexibility that pseudolites can offer, (PL) and a receiver (Rec), which can essentially receive pseudolites can be combined with other sensors such as the signals from all the transmitters including the one INS. In contrast to satellite/pseudolite-based positioning inside the transceiver itself. This hardware design enables systems, INS is self-contained and autonomous. Thus, the cancellation of both transmitter and receiver clock INS systems are independent of any external signals. errors without using a separate reference station, which is However, one of main drawbacks of INS, when operated normally required for typical differential satellite as a stand-alone system, is the time-dependent growth of positioning. systematic errors. GPS measurements are typically used to calibrate INS systematic errors. GPS signals might be obstructed for extended time periods under difficult operational conditions, during which the performance of integrated GPS/INS systems may degrade rapidly. This issue can be addressed by the inclusion of pseudolite signals. An integrated GPS/INS/pseudolite or INS/pseudolite system would be able to improve system performance under a wide variety of poor operational environments. Such an integration concept has been proposed and tested (Wang et al., 2001c, Grejner- Brzezinska et al., 2002), demonstrating the feasibility and potential for a GPS/INS/PL system. Fig. 5 Double differencing with transceivers (LeMaster, 2002) Extensive simulation studies have demonstrated that an Pseudolite-only postioning can also be based on the so- integrated PL/INS system can provide a stable and high called inverted positioning concept (Raquet et al., 1995). precision navigation solution for indoor applications (Lee In such a positioning scenario a reference pseudolite, a et al., 2002). The reliability and accuracy of an integrated user/mobile pseudolite and four or more receivers are PL/INS system is dependent on the geometric distribution needed. Similar to GPS relative positioning, the double- of the pseudolites deployed within the system. Figure 6 differenced measurements between pseudolites and shows the DOP values for two different pseudolite receivers can be formed to remove most of the systematic geometric scenarios, indicating a significant difference in errors, such as transmitter and receiver clock errors. The the geometric strength. As shown in Figure 7, the better pseudolite geometry will lead to better performance

Wang: Pseudolite Applications in Positioning and Navigation 53 within an integrated PL/INS system. It is noted that the Pseudolite hardware vertical component can be even more accurate than the All commercial pseudolite products are currently using horizontal components. This is due to the use of GPS L1 frequency. Operating such pseudolites requires pseudolite measurements from low elevation angles. care so that pseudolite signals do not jam or interfere with Overall, these simulation results show that an integrated the operation of nearby GPS receivers. Although pulsing PL/INS system can achieve centimetre-level accuracy pseudolite signals can reduce the potential interference even in indoor operating environments (ibid, 2002). with GPS signals, there could be some other frequency 15 choices suitable for pseudolite applications. To select RDOP RHDOP 10 optimal pseudolite frequencies, careful considerations RVDOP Mean RDOP : 4.7 should be given to such factors as hardware 5 implementation, frequency allocation/licensing, 0 ambiguity resolution, multipath mitigation, as well as 5.04 5.041 5.042 5.043 5.044 5.045 5.046 5.047 5.048 5.049 5 x 10 integration with GPS/GLONASS/GALILEO or even Value 5 mobile phone signals for mobile location applications. 4 RDOP RHDOP Mean RDOP : 2.1 3 RVDOP Mobile phone signal transmitters can also be considered

2 as ‘pseudo-satellites’ in some sense (Zhao, 2000). 1 Although there are several solutions proposed to the 0 5.04 5.041 5.042 5.043 5.044 5.045 5.046 5.047 5.048 5.049 ‘near-far’ problem, new solutions are still emerging that 5 GPS second x 10 may offer more flexibility and improved performance in Fig. 6 DOP values for two different pseudolite scenarios pseudolite positioning.

PL/INS Integation Pseudolite receivers 0.6 0.3 Experiments with current pseudolite and receiver 0 hardware have revealed some characteristics of -0.3 Standard Dev. ∆ North Coordinates pseudolite signals, which need to be considered in the -0.6 ∆ N : 20.1mm ∆ East Coordinates ∆ E : 20.9mm -0.9 ∆ Vertical Coordinates receiver tracking loops (Ford et al., 1996; Biberger et al., ∆ V : 30.7mm -1.2 2001). To develop a robust pseudolite tracking receiver,

Position Difference(m) 5.04 5.041 5.042 5.043 5.044 5.045 5.046 5.047 5.048 5.049 5 x 10 more investigations are required to gain insights into the 0.6 pseudolite signal propagation and reception under a 0.3 variety of operating conditions, for example, high 0 -0.3 dynamics and severe multipath. Based on the ultra-tight Standard Dev. North Coordinates -0.6 ∆ N : 18.8mm ∆ GPS/INS/PL or INS/PL integration concept, a new ∆ E : 13.0mm ∆ East Coordinates -0.9 ∆ V : 10.8mm ∆ Vertical Coordinates receiver design with an inertial aid may improve the -1.2 Position Difference(m) 5.04 5.041 5.042 5.043 5.044 5.045 5.046 5.047 5.048 5.049 signal tracking and enhance the reliability of the 5 GPS second x 10 positioning solutions. Software receiver architecture Fig. 7 Differences between the reference and computed coordinates for appears to be a good platform for such developments. the PL/INS simulation (Lee et al., 2002) Multipath Although these simulations demonstrate the potential of GPS/INS/PL or PL/INS integration, there have been In pseudolite applications, particularly for indoor some concerns about the systematic biases in pseudolite positioning, multipath is a major concern. In static measurements (e.g., Wang et al., 2001b;c; Dai et al., positioning, pseudolite multipath biases appear to be 2001). constant. In kinematic mode, however, the possible biases are most likely randomised and thus are difficult to deal with. This issue may be addressed by using appropriate transmitting and receiving antennas, robust tracking 4 Challenging Issues in Pseudolite Applications techniques, as well as sensor integration. Helical transmitting antennas are commonly used to mitigate From the positioning point of view, a pseudolite is just multipath, whilst various well-designed GPS receiver like a satellite on the ground. However, different antennas can also contribute to the reduction of multipath. locations of spaceborne satellites and ground-based radio Given the complexity of the multipath issue associated transmitters will have a variety of implications for with spread spectrum ranging, there will be continued positioning performance. While pseudolites can offer research on multipath mitigation techniques. In another great flexibility in terms of geometry and signal development, new radio-location systems based on the availability, the small separation between pseudolites and Ultra Wide Band (UWB) technology are being proposed users may cause, among others, a ‘near-far’ problem in (http://www.uwb.org). It is expected that the new UWB signal tracking, multipath, and tropospheric delay errors location systems may more efficiently mitigate multipath in modelling. Therefore, there are challenging issues in pseudolite applications, which need to be addressed.

54 Journal of Global Positioning Systems in indoor positioning, as UWB radio signals are sources. The statistics of the errors relating to signal transmitted as very short discrete pulses. propagation need to be investigated. Pseudolite synchronisation Unlike GPS satellites, pseudolites are usually equipped 5 Concluding Remarks with low cost clocks, which are not accurate enough to synchronize the sampling time between the reference and Global navigation satellite systems are expected to play user receivers in a differential positioning mode. Further an increasingly important role in the worldwide more, in a single positioning mode, synchronization of all geospatial information infrastructure. During the past the pseudolites used within a positioning system is decade GPS has been the driving force behind numerous critical, which has been one of technical challenges in position/location sensitive applications, such as car pseudolite-only applications. Various techniques have navigation, mobile phone location, and many others. been proposed to deal with these problems (e.g., Cobb, Perhaps the most important contribution that GPS has 1997; Kee et al., 2000; Söderholm et al., 2001). It is made is to raise the awareness of location information. highly desirable to develop synchronizing strategies Ultimately we will expect to have precise and reliable suitable for various operating environments. If the location information for any object in real-time anywhere synchronization errors are reduced to the noise level of and at any time. Location technologies will be an the carrier phases, single-differenced integer carrier phase indispensable part of many emerging areas of business, ambiguities can be resolved and thus, centimetre-level such as Location-Based Services (LBS) and future positioning accuracy can be achieved using just one personal navigation devices, “smart highway” systems, receiver. and so on. However, current satellite-based positioning Modelling errors systems cannot meet all the requirements for location information, which include accuracy, reliability, integrity, In pseudolite positioning, major error sources, such as coverage and availability. Therefore, new location pseudolite location errors, multipath and troposheric technologies are desperately needed to enhance, even delays, need to be treated carefully. In a differential replace under certain circumstances, the satellite positioning mode, pseudolite location errors can be positioning systems. As a by-product of GPS system doubled in the differenced measurements with an development, the pseudolite concept and technology has unfavourable geometry (e.g., Hein et al., 1997; Dai et al., demonstrated great potential in meeting such a need. 2002). Therefore, pseudolite locations should be precisely determined. Acknowledgements Multipath might be ‘amplified’ in the single-or double- differenced measurements. Although multipath should be The author would like to thank Prof. Chris Rizos for his carefully dealt with by using appropriate valuable comments on this paper. Thanks are extended to other hardware/firmware, optimal pseudolite-receiver geometry members of the Pseudolite Working Group (Michalson & Progri, 2000) and modelling procedures (http://www.gmat.unsw.edu.au/pseudolite) within the can also contribute to the reduction of multipath. International Association of Geodesy for many discussions and the accumulated references they have provided on pseudolite It has been estimated that tropospheric delays in the research. pseudolite measurements may reach up to 320ppm with standard meteorological parameters, and even 600ppm under extreme weather conditions (Dai et al., 2001a). To References reliably model the tropospheric delays, precise Altmayer C., Martin S., & Theil S. (1998) Autonomous measurements of atmospheric pressure, temperature and onboard orbit and attitude control of geostationary humidity should be available. Barltrop et al. (1996) have satellites using pseudolites, Proceedings of US Institute of proposed an adaptive tropospheric delay estimation Navigation GPS-98, Nashville, TN, USA, 15-18 method, in which troposheric delays are considered as September, 1565-1574. additional unknown parameters in the positioning solutions. More investigations are required to evaluate the Barltrop K.J., Stafford J.F. & Elrod B.D. (1996) Local DGPS with pseudolite augmentation and implementation accuracy and reliability of such estimated delays. considerations for LAAS, Proceedings of US Institute of In the real world, precisely modelling range Navigation GPS-96, Kansas City, Missouri, 17-20 Sept., measurements from radio signals is a challenge. Under 449-459. difficult operating conditions, for example, when signals Barnes J., Wang J., Rizos C., & Tsujii T. (2002) The propagate within buildings and penetrate through walls performance of a pseudolite-based positioning system for (Rappaport & Sandhu, 1994; Pahlavan et al., 1998; deformation monitoring. Proceedings of the 2nd Symp. on Peterson et al., 2000), there are potentially many error Geodesy for Geotechnical & Structural Applications, Berlin, Germany, 21-24 May, 326-327.

Wang: Pseudolite Applications in Positioning and Navigation 55

Bartone C. G. & Kiran S. (2001) Flight test results of an Galijan R.C. (1996) Analysis and simulation of a candiate integrated wideband airport pseudolite for the local area deployment geometry and characteristics of pseudolites augmentation system, Navigation, 48(1): 35-48. with a tunnel, Proceedings of US Institute of Navigation GPS-96, Kansas City, 17-20 September, 527-533. Beser J. & Parkinson B.W. (1982) The application of NAVSTAR differential GPS in the civilian community, Grejner-Brzezinska D., Yi Y., Wang J. (2002) Design and Navigation, 29(2): 107-136. navigation performance analysis of an experimental GPS/INS/PL system, 2nd Symp. on Geodesy for Biberger R.J., Hein G.W., Eissfeller B., Oehler V., Schueler T. Geotechnical & Structural Applications, Berlin, Germany, (2001) Pseudolite signal creeping on conducting 21-24 May, 452- 461. surfaces, Proceedings of US Institute of Navigation GPS- 2001, Salt Lake City, Utah, 11-14 September, 1238-1244. Harrington R.L. & Dolloff J.T. (1976) The inverted range: GPS user test facility, IEEE PLANS’76, San Diego, Brown A.K. (1992) A GPS precision approach and landing California, 1-3 Nov., 204-211. system, Proceedings of US Institute of Navigation GPS-92, Albuquerque, New Mexico, 16-18 Sept., 373-381. Hein G.W. (2002) Personal communications. Cobb H.S. (1997) GPS pseudolites: Theory, design, and Hein G.W., Eissfeller B., Werner W., Ott B., Elrod B.D., applications, PhD Dissertation, Stanford University. Barltrop K.J. & Stafford J.F. (1997) Practical investigation on DGPS for aircraft precision approaches Cohen C.E., Pervan B.S., Cobb, H.S., Lawrence D.G., Powell augmented by pseudolite carrier phase tracking, J.D. & & B.W. Parkinson (1993) Real time cycle Proceedings of US Institute of Navigation GPS-97, Kansas ambiguity resolution using a pseudolite for precision City, Missouri, 16-19 Sept., 1851-1860. landing of aircraft with GPS, The Second International Symposium on Differential Satellite Navigation Systems Holden T. & Morely T. (1997) Pseudolite augmented DGPS DSNS’93, Amsterdam, The Netherlands, March 30-April for land applications, Proceedings of US Institute of 2, 171-178. Navigation GPS-97, Kansas City, Missouri, 16-19 Sept., 1397-1404. Corazzini T. & How J.P. (1999) Onboard Pseudolite Augmentation System for Relative Navigation, ITT (2002) 746th GPS Test Range, GPS World, February, Proceedings of US Institute of Navigation GPS-99, http://www.gpsworld.com. Nashville, TN, Sept. 1999, pp. 1559-1568 Kalafus R.M., Van Dierendock A.J. & Pealer N. (1986) Special Dai L., Wang J., Tsujii T., & Rizos C. (2001a) Pseudolite Committee 104 recommendations for differential GPS applications in positioning and navigation: Modelling service. Global Positioning System (red book), The and geometric analysis, Int. Symp. on Kinematic Systems Institute of Navigation, Vol. I, 101-126. in Geodesy, Geomatics & Navigation (KIS2001), Banff, Canada, 5-8 June, 482-489. Kee C., Jun H., Yun D., Kim B., Kim Y., Parkinson B.W., Langestein T., Pullen S. & Lee J. (2000) Development of Dai L., Wang J., Rizos C. & Han S. (2001b) Pseudo-satellites indoor navigation system using asynchronous applications in deformation monitoring, GPS Solutions, pseudolites, US Institute of Navigation GPS-2000, Salt 5(3), 80-87 Lake City, Utah, 19-22 Sept, 1038-1045. Dai L., Wang J., Tsujii T., & Rizos C. (2002) Inverted Klein D. & Parkinson B.W. (1986) The use of pseudo-satellites pseudolite positioning and its applications, Accepted for for improving GPS performance, Global Positioning publication Survey Review. System (red book), Vol III, Institute of Navigation, 1986, 135-146. Dovis F., Kandus G., Magli E. & Olmo G. (2000) Integration of stratospheric platforms within the GNSS2 system, Lee H.Y., Wang J., & Rizos C. (2002) Kinematic positioning Proceedings of GNSS-2000, Edinburgh, U.K., 1-4 May. with an integrated GPS/pseudolite/INS, 2nd Symp. on CD-ROM proc. Geodesy for Geotechnical & Structural Applications, Berlin, Germany, 21-24 May, 314-325. Elrod B.D. & Van Dierendonck A.J. (1996) Pseudolites. In: B.W. Parkinson & J.J. Spilker (eds.), Global Positioning LeMaster E.A. (2002) Self-Calibrating Pseudolite Arrays: System: Theory and Applications (Vol. II), American Theory and Experiment, Ph.D. Thesis, Stanford Institute of Astronautics, Washington D.C., 51-79. University. Ford T., Neumann J., Toso N., Petersen W., Anderson C., LeMaster E. A. & Rock S. (1999) Mars exploration using self- Fenton P., Holden T. & Barltrop K. (1996) HAPPI – A calibrating pseudolite arrays, Proceedings of US Institute high accuracy pseudolite/GPS positioning integration, of Navigation GPS-99, Nashville, Tennessee, 14-17 Sept., Proceedings of US Institute of Navigation GPS-96, Kansas 1549-1558. City, Missouri, 17-20 Sept., 1719-1728. Madhani P.H., Axelrad P., Krumvieda K. Thomas J. (2001) Galijan R.C. and Lucha G.V. (1993) A suggested approach for Mitigation of the near-far problem by successive augmenting GNSS Category III approaches and interference cancellation, Proceedings of US Institute of landings: the GPS/GLONASS and Glonass pseudolite Navigation GPS-2001, Salt Lake City, Utah, 11-14 Sept, system, Proceedings of US Institute of Navigation GPS-93, 148-154. Salt Lake City, Utah, 22-24 September, 157-160. Meng X., Roberts G.W., Dodson A.H., Cosser E., Noakes C. (2002) Simulation of the effects of introducing pseudolite

56 Journal of Global Positioning Systems

data into bridge deflection monitoring data, Proceedings Tsujii T., Rizos C., Wang J., Dai L., Roberts C. & Harigae M. of 2nd Symp. on Geodesy for Geotechnical & Structural (2001) A navigation/positioning service based on Applications, Berlin, Germany, 21-24 May, 372-381. pseudolites installed on stratospheric airships, 5th Int. Symp. on Satellite Navigation Technology & Applications, Michalson M.R. & Progri .I.F. (2000) Assessing the accuracy Canberra, Australia, 24-27 July, CD-ROM. of underground positioning using pseudolites, Proceedings of US Institute of Navigation GPS-2000, Salt Tuohino J.L., Farley M.G. & James R.R. (2000) Military Lake City, Utah, 19-22 Sept., 1007-1005. pseudolite flight test results, Proceedings of US Institute of Navigation GPS-2000, 19-22 September, Salt Lake Morley T. & Lachapelle G. (1998) Pseudolite augmentation City, UT, 2079-2088. for OTF ambiguity resolution in shipborne mode, Journal of Surveying Engineering, 124(1): 26-39. Van Dierendonck A.J., Elrod B.D. and Melton W.C. (1989) Improving the integrity, availability and accuracy of GPS O’Keefe K., Sharma J., Cannon M.E. & Lachapelle G. (1999) using pseudolites, Proceedings of NAV’89, London, UK, Pseudolite-based inverted GPS concept for local area Oct 17-19, 1989, paper 32. positioning, Proceedings of US Institute of Navigation GPS-99, Nashville, Tennessee, 14-17 Sept., 1523-1530. Verhagen S. (2001) Ambiguity resolution and success rates with an integrated GNSS - pseudolite positioning system, Rappaport T.S. & Sandhu S. (1994) Radio-wave propagation Proceedings of US Institute of Navigation GPS 2001, Salt for emerging wireless personal-communication systems, Lake City, Utah, 11-14 September 1436-1478. IEEE Antennas and Propagation Magazine, 36(5):14-23. Verhagen S. (2002) Internal and external reliability of an Raquet J., Lachapelle J., Qui W., Pelletier C., Nash A., Fenton integrated GNSS – pseudolite positioning system, P. & Holden T. (1995) Development and testing of a Proceedings of 2nd Symp. on Geodesy for Geotechnical & mobile pseudolite concept for precise positioning, Structural Applications, Berlin, Germany, 21-24 May, 314- Proceedings of US Institute of Navigation GPS-95, Palm 325. Springs, California, 12-15 Sept., 817-825. Wang J. (2000) An approach to Glonass ambiguity resolution, Pachter M. & Mckay J.B. (1998) Geometry optimization of a Journal of Geodesy, 74(5), 421-430. GPS-based navigation reference system, Navigation, 44(4), 457-470. Wang J., Rizos C., Stewart M.P., & Leick A. (2001a) GPS and Glonass integration: Modelling and ambiguity resolution Pahlavan K., Krishnamurthy P. and Beneat J. (1998) Wideband issues, GPS Solutions, 5(1), 55-64. radio propagation modelling for indoor geolocation applications, IEEE Coummunication Magazine, April, 60- Wang J., Tsujii T., Rizos C., Dai L. & Moore M. (2001b) GPS 65. and pseudo-satellites integration for precise positioning. Geomatics Research Australasia, 74:103-117. Parkinson B.W. & Fitzgibbon K.T. (1986) Optimal locations of pseudolites for differential GPS, Navigation, 33(4), 259- Wang J., L. Dai, T. Tsujii, C. Rizos, D. Grejner-Brzezinska & 283. C.K. Toth (2001c) GPS/INS/Pseudolite integration: Concepts, simulation and testing, Proceedings of US Pervan B., Cohen C.E. & Parkinson B.W. (1994) Integrity Institute of Navigation GPS-2001, Salt Lake City, Utah, monitoring for precision approach using kinematic GPS 11-14 September, 2708-2715 and a ground-based pseudolite, Navigation, 41(2), 159- 174 Wawrzyniak G., Lightsey E.G., Key K.W. (2001) Ground experimentation of a pseudolite-only method for the Peterson B.B., Kmiecik C.G., Nguyen H. & Kaspar B. (2000) relative positioning of two spacecraft, Proceedings of US Indoor geolocation system operational test results, Institute of Navigation GPS-2001, Salt Lake City, Utah, Navigation, 47(3):157-166. 11-14 September 1436-1478 Progri I.F. & Michalson W.R. (2001) An Alternative Approach Weiser M. (1998) Development of a carrier and C/A-code to Multipath and Near-Far Problem for Indoor based pseudolite system, Proceedings of US Institute of Geolocation Systems, Proceedings of US Institute of Navigation GPS-98, Nashville, USA, September, pp. 1465- Navigation GPS-2001, Salt Lake City, Utah, 11-14 1475. September 1434-1467 Zhao Y. (2000) Mobile phone location determination and its Söderholm S., Juhola T., Saarnimo T., Karttunen V. (2001), impact on intelligent transportation systems, IEEE Indoor navigation using a GPS receiver, Proceedings of Transactions on Intelligent Transportation Systems, US Institute of Navigation GPS-2001, Salt Lake City, 1(1):55-64. Utah, 11-14 September, 1479-1486 Zimmerman K.R. (1996) Experiments in the use of Global Stansel T. A. Jr. (1986) RTCM SC-104 Recommended Positioning System for space vehicle rendezvous, PhD pseudolite signal specification, Global Positioning thesis, Standford University. System, Vol. 3, The US Institute of Navigation, 1986, 117- 134. Zimmerman K.R., Cohen C.E., Lawrence D.G., Montgomery P.Y., Cobb H.S. & Melto W.C. (2000) Multi-frequency Stone J.M., LeMaster E.A., Powell J.D. & Rock S. (1999) GPS pseudolites for instantaneous carrier ambiguity Pseudolite Transceivers and Their Applications, resolution, Proceedings of US Institute of Navigation Proceedings of US ION National Technical Meeting, San GPS-2000, Salt Lake City, Utah, 19-22 Sept., 1024-1030 Diego, California, 25-27 January, 415-424.

Journal of Global Positioning Systems (2002) Vol. 1, No. 1: 58-60

An Overview of Atmospheric Radio Occultation

T. P. Yunck Jet Propulsion Laboratory, California Institute of Technology Pasadena, California 91109

Received: 30 April 2002 / Accepted: 30 April 2002

know well, such measurements must be continuous, comprehensive, synoptic. We therefore need many Biography transmitters and receivers aloft at once, densely sampling the global atmosphere every few hours. Thomas P. Yunck, holds a bachelor's degree in electrical In the late 1980s, a group at JPL proposed observing GPS engineering from Princeton University and a Ph.D. in signals from space to make atmospheric soundings, as systems and information science from Yale University. shown in Fig. 1. Briefly, the observed Doppler shift in the Since 1978 he has been with the Jet Propulsion GPS signal induced by atmospheric bending permits Laboratory, California Institute of Technology, where he accurate estimation of the atmospheric refractive index. currently manages the GPS Observatories Office. At JPL, From that one can retrieve, in sequence, profiles of the Dr. Yunck has been involved in the development of radio atmospheric density, pressure, and temperature (or, in the metric techniques for spacecraft navigation and for a lower troposphere, water vapor) with high accuracy (<1 variety of related science pursuits. For the past 15 years Kelvin in temperature) and a vertical resolution of a few he has managed the development of technologies to hundred meters. Figure 2 shows the predicted accuracy of employ the signals from GPS for high precision Earth atmospheric temperature profiles as a function of altitude. science and remote sensing. His current work focuses on In the lower part of the troposphere, the uncertainty in the development of spaceborne GPS systems for water vapor content, particularly in the tropics, leads to a applications in geodesy, atmospheric sounding, and large error in the recovered temperature. In that region, ionospheric imaging. since it is water vapor that is of greater consequence in weather modeling, it becomes advantageous to adopt nominal temperature lapse rates and instead recover water vapor profiles. When a radio signal passes through the atmosphere its phase is perturbed in a manner related to the refractivity along the ray path. Measurements of the phase perturbations can reveal the refractivity, from which one can then derive such quantities as atmospheric density, pressure, temperature, moisture, geopotential heights, and winds. This general technique is known as atmospheric radio occultation. The probing of planetary atmospheres by radio occultation dates to the early 1960s when Mariners 3 and 4, viewed from Earth, passed behind Mars. While radio Fig. 1 Atmospheric temperature profiling by GPS occultation occultation has probed planets and moons throughout the solar system, it has as not found operational application A single satellite can recover more than 500 profiles each to Earth, for two reasons. First, the observation requires day, distributed almost uniformly around the globe; a both a radio source and a suitable receiver off the planet, large constellation would recover many thousands of pro- outside the atmosphere; seldom have we had such files, which could one day have a profound impact on matched pairs in Earth orbit. Second, to be of use in both long term climatological studies and short term studying Earth's atmosphere, the nature of which we weather modeling.

Yunck: An Overview of Atmospheric Radio Occultation 59

In the early 1990’s, a group led by the University systems will be further miniaturized and we will see Corporation for Atmospheric Research (UCAR) dedicated constellations of dozens or more free-flyers, succeeded in obtaining sponsorship from the U.S. each with a mass of a few kilograms (Fig. 4). There is National Science Foundation for a low-cost much to be said for this view. Consider that the next demonstration experiment called GPS/MET. UCAR and generation of miniature occultation receiver will also JPL converted a low cost geodetic ground receiver to fly provide: in space and acquire the occultation data. • Real time onboard position, velocity, attitude, and timing • All onboard spacecraft computation and control • Uplink extraction and command interpretation

Fig. 2 Approximate occultation temperature accuracy vs altitude.

Fig. 4 Concept for a pilot constellation of spaceborne GPS sensorcraft for Earth science. Equipped with a cell-phone chip, each orbiter will send its data directly to a computer which, within minutes, will generate and distribute finished science products. No special ground systems are needed for tracking or data downlink, and operations crews would be eliminated. These modest components can in principle be packaged in a tiny spacecraft weighing just a few kilograms. Fig. 3 GPS temperature profile compared with two radiosondes. Another view holds that these same virtues will make the Figure 3 shows a typical GPS/MET temperature profile, GPS sensor irresistible as an add-on to constellations along with nearby radiosonde measurements for com- already planned. The miniature receiver/processor can parison. Recently, five new spacecraft have been assume the functions of several discrete spacecraft sub- launched into low Earth orbit carrying NASA occultation systems, lowering mass, power, and cost in a cascade of receivers: CHAMP (Germany), SAC-C (Argentina), economies, while providing a valuable new science GRACE (2 spacecraft, US/Germany), and IOX (US). Still dimension. The cost of adding GPS science to a suitable others are being planned by the US, Europe, Japan, constellation would be small. Brazil, and Taiwan. In the immediate future, the latter scenario may be more While these missions will do much to advance likely. The investment needed to realize the sensorcraft spaceborne GPS science, they will not in themselves concept is not trivial and science agencies tend to proceed establish an ongoing presence of large numbers of GPS with caution. The vision of great constellations has sensors in earth orbit. It is the hope of the growing GPS instead been taken up by industry, which has already Earth science community that a permanent constellation deployed several commercial constellations, and to a of a dozen or more occultation sensors may emerge in the lesser extent by the military. Together these near future. constellations will comprise hundreds of spacecraft wonderfully suited to GPS occultation sensing. There has been much speculation on how this might Whichever direction the future takes it promises to be a unfold. In one view, the model of a dedicated constellation will be taken to its logical extreme: flight

60 Journal of Global Positioning Systems rewarding one for those engaged in exploring our System, J. Geophys. Res., 102, No. D19, pp. 23429-23465, complex Earth system. 1997. Rocken, C., R. Anthes, M. Exner, R. Ware, D. Feng, M. Gorbunov, B. Herman, D. Hunt, Y.-H. Kuo, W. Schreiner, Bibliography S. Sokolovskiy, and X. Zou: Verification of GPS/MET Data in the Neutral Atmosphere, Journal of Geophysical Lee, L.-C., C. Rocken and R. Kursinski (eds.): Applications of Research, 102, pp. 29,849-29,866, 1997. Constellation Observing System for Meteorology, Melbourne, W. G. et al.: The Application of Spaceborne GPS Ionosphere & Climate, Springer-Verlag, Hong Kong , 384 to Atmospheric Limb Sounding and Global Change pp, 2000. Monitoring, JPL Publication 94-18, Jet Propulsion Kursinski, E. R., G. A. Hajj, J. T. Schofield, R. P. Linfield and Laboratory, California Institute of Technology, Pasadena, K. R. Hardy: Observing the earth's atmosphere with radio 147 pp., 1994. occultation measurements using the Global Positioning

Journal of Global Positioning Systems (2002) Vol. 1, No. 1: 61-63

The Contribution of GPS Flight Receivers to Global Gravity Field Recovery

Peter Schwintzer, Christoph Reigber GeoForschungsZentrum Potsdam, Germany

Received: 15 May 2002 / Accepted: 15 May 2002

Rogue Space Receiver (TRSR-2) is provided by NASA and manufactured at NASA's Jet Propulsion Laboratories Biography (JPL) [Kuang et al., 2001]. The purpose of this instrument is to allow a recovery of CHAMP's trajectory Christoph Reigber, Prof. Dr.-Ing. Dr.-Ing. E.h, Director with an uncertainty of only a few centimetres. The of Division 1 'Kinematics and Dynamics of the Earth' of receiver acquires up to 12 GPS satellites simultaneously GeoForschungsZentrum Potsdam, is in particular and measures dual-frequency carrier phases and pseudo- engaged in all aspects of satellite geodesy and its relation ranges at a rate of 10 s. Monitoring CHAMP's orbit by to geotectonics, Earth rotation, Earth gravity field, GPS allows the observation of gravity induced orbit oceanography and atmosphere/ionosphere. perturbations which then are analysed to map the global structure of the Earth's gravitational field [Reigber et al., He is Director of the CHAMP mission, Co-Principal 1999] (Figure 1). Investigator of the GRACE mission and Chair of the Coverning Board of the international GPS Service. Peter Schwintzer, Dr.-Ing., Head of Section 1.3, 'Gravity Field and Figure of the Earth' within Division 1 of GeoForschungsZentrum Potsdam, is engaged in global gravity field modelling from space and its geophysical application. He is the Science Data System manager of the CHAMP mission.

On July 15, 2000, the German geoscientific satellite CHAMP (CHAllenging Mini-satellite Payload) was launched into an almost circular, near-polar orbit with an initial altitude of 454 km, slowly decreasing to 300 km during the expected lifetime of five years. The CHAMP Fig. 1 High-low GPS-CHAMP satellite-to-satellite tracking for gravity mission is conducted since the beginning under full field recovery responsibility of GeoForschungsZentrum Potsdam with Earth gravity field recovery from observed satellite orbit participation of the German Centre for Aerospace perturbations has been applied since the beginning of the (DLR).The mission is funded by the German Ministry of space age in the late 1950s and evolved to long- Education and Research (cf. GFZ website http://op.gfz- wavelength gravity field models which today resolve potsdam.de/CHAMP). spatial features in the gravity field with half wavelengths larger than 500 km at the Earth's surface. The models For the first time a satellite in such a low altitude is which were generated prior to the launch of CHAMP equipped with a GPS receiver. The 2nd generation Turbo exploited mainly ground-based camera, microwave and

62 Journal of Global Positioning Systems laser tracking data from some tens of satellites at different The sum of these mission characteristics already led to a altitudes and orbit inclinations [Biancale et al., 2000]. break-through in the determination of the long- With CHAMP it becomes for the first time possible to wavelength gravitational field [Reigber et al., 2002]. The derive a global gravity field model from orbit ability to achieve, thanks to uninterrupted GPS space- perturbations gathered over a short time interval of a few based tracking, accurate gravity field solutions from only months from one satellite only (Figure 2). Moreover, the a short observation time interval also opens the resulting model is up to four times more accurate than possibility to study non-tidal temporal gravitational field what has been achieved with the earlier multi-satellite variations. These are mainly due to seasonal, interannual solutions and multi-year tracking records. Geodesy, and long-term mass redistributions in and among the Oceanography and Geophysics benefit from the advanced Earth's atmosphere, hydrosphere and cryosphere. The knowledge of the Earth's gravity field. observation of these phenomena which are relevant for climate studies is in particular the objective of the GRACE (Gravity Recovery And Climate Experiment) mission. The two satellites of the GRACE constellation were launched on March 17, 2002, for a 5 years' mission. The GRACE satellites are part of NASA's Earth's System Science Pathfinder project. The German Center for Aerospace (DLR) participates in the mission. The science processing system is chaired by the Center for Space Research (CSR) of Texas University in Austin with a distribution of work between CSR, JPL and GFZ [Tapley and Reigber, 2001]. The two GRACE satellites, flying one after the other at a distance of about 220 km in an initial altitude of 500 km and in a near-polar orbit, are similar to CHAMP. Both carry a GPS receiver and an accelerometer, but with an increased resolution. The new element of the GRACE mission is the K-Band Ranging System (KBR) which measures the dual one-way range between both satellites Fig. 2 Global gravity field model from three months of CHAMP data with a precision of about 5 µm. By measuring gravity- only induced relative distance variations between the two satellites, the resolution in global gravity field recovery The advantages of the CHAMP mission with respect to from space can be very likely extended from about 500 all former geodetic gravity missions are the following: (1) km to 150 km (half-wavelength) with a gain in accuracy Orbit configuration - The effect of the attenuation of the by one to two orders in magnitude compared to the gravitational signal with altitude is minimized due to the present knowledge (CSR website http://www.csr.utexas low orbit altitude, and there is no restriction in ground .edu/grace, and GFZ website http://op.gfz-potsdam.de track coverage thanks to the almost polar orbit. (2) GPS /grace). GRACE also relies on the exploitation of GPS receiver – The on-board GPS receiver allows continuous high-low satellite-to-satellite tracking data for the tracking by up to 12 GPS satellites simultaneously restitution of the long-wavelength part of the gravitational compared to one-dimensional ground-based tracking of spectrum. only short orbit pieces during satellites passes. (3) Accelerometer – CHAMP experiences at its low altitude The third satellite in the sequence of recent dedicated enhanced accelerations due to air drag. These non- gravity satellite missions will be GOCE (Gravity field gravitational orbit perturbations have to be accounted for and steady-state Ocean Circulation Explorer). GOCE is when using the GPS observed overall orbit perturbations planned to be launched in 2006 and was selected as the for gravity field recovery. The on-board three axes first Core Mission within the Earth Explorer Programme accelerometer, provided by the French space agency of the European Space Agency [ESA 1999]. The payload CNES and manufactured by the French company of GOCE will consist of a GPS/GLONASS receiver, ONERA directly measures the vector of non-gravitational again for resolving the long-wavelength gravity field, and accelerations, i.e. air drag plus direct and indirect solar an assembly (gradiometer) of six three-axes radiation pressure [Touboul et al., 1999]. These accelerometers to measure in-orbit gravity gradients. For measurements replace air density models which are of the first time gravity field recovery from space will not be insufficient accuracy and temporal resolution. The based purely on the analysis of orbit perturbations. orientation of the accelerometer's axes is known from two GOCE will fly in an extremely low orbit (250 km star camera. altitude) which is permanently maintained by ion- thrusters compensating for air-drag (drag-free concept).

Schwintzer et al: The Contribution of GPS Flight Receivers 63

The relatively short mission with twelve months selection, in The four candidate Earth explorer core observation time, aims at an ultimately high and accurate missions, SP-1233 (1), Noordwijk, The Netherlands, 1999. resolution of the gravity field down to half wavelengths Kuang, D., Bar-Server, Y., Bertiger, W., Desai, S., Haines, B., below 100 km. By this, the requirement of the Iijima, B., Kruizinga, G., Meehan, Th., and L. Romans, oceanographers for a high-resolution precise geoid shall Precise Orbit Determination for CHAMP using GPS be fulfilled. This is needed as a physical reference surface Data from BlackJack Receiver, in 2001 ION National for the determination of the global ocean circulation Technical Meeting Proceedings, Session E1: Scientific pattern with satellite altimetry. Applications, Timing, and Frequency, Long Beach, California, January, 2001. The combination of GPS high-low satellite-to-satellite Reigber, Ch., Balmino, G., Schwintzer, P., Biancale, R., Bode, tracking with accelerometry, a low-low intersatellite link A., Lemoine, J.-M., König, R., Loyer, S., Neumayer, H., and/or a gradiometer on low Earth orbiting platforms Marty, J.-Ch., Barthelmes, F., Perosanz, F., and S.Y. Zhu, provides an excellent tool for mapping the Earth's gravity A High-Quality Global Gravity Field Model from field homogeneously from space with ever increased CHAMP GPS Tracking Data and Accelerometry accuracy and resolution over the globe and in time. (EIGEN-1S), Geophys. Res. Letters, 29(14), 10.1029/ 2002GL015064, 2002. Reigber, Ch., Schwintzer, P., and H. Lühr, The CHAMP Bibliography geopotential mission, Boll. Geof. Teor. Appl., 40, 285- 289, 1999. Biancale, R., Balmino, G., Lemoine, J.-M., Marty, J.-C., Moynot, B., Barlier, F., Exertier, P., Laurain, O., Gegout, Tapley, B.D., and Ch. Reigber, The GRACE mission: status P., Schwintzer, P., Reigber, Ch., Bode, A., König, R., and future plans, Eos Trans AGU 82 (47), Fall Meet. Massmann, F.-H., Raimondo, J.-C., Schmidt, R., and S.Y. Suppl., G41 C-02, 2001. Zhu, A New global Earth's Gravity Field Model from Touboul, P., Willemenot, E., Foulon, B., and V. Josselin, Satellite Orbit Perturbations: GRIM5-S1, Geophys. Res. Acclerometers for CHAMP, GRACE and GOCE space Letters, 27, 3611-3614, 2000. missions: synergy and evolution, Boll. Geof. Teor. Appl., European Space Agency, Gravity Field and Steady-State 40, 321-327, 1999. Ocean Circulation Mission (GOCE), Report for mission

Journal of Global Positioning Systems (2002) Vol. 1, No. 1: 64-65

Ocean Remote Sensing with GPS

Cinzia Zuffada Jet Propulsion Laboratory California Institute of Technology Pasadena, California 91109

Received: 5 May 2002 / Accepted: 16 June 2002

A GPS receiver in low-Earth orbit (LEO) with an antenna pointed toward the Earth’s surface can, in principle, track Bibliography about 10 GPS reflections simultaneously, therefore providing a coverage that is an order of magnitude denser Cinzia Zuffada, Senior Member Technical Staff in the than nadir-viewing altimeters. For example, the reflection Earth Orbiter Systems Group, Jet Propulsion Laboratory ground tracks of a satellite at the altitude of 400 km (JPL). Her work at JPL has been focused on the use of would cover the Earth nearly uniformly in just 1 day, GPS for remote sensing of the atmosphere and the ocean. with at most about 75 km across-track separation. Such She has been managing the research in GPS altimetry at dense coverage can be translated into a higher temporal JPL for the past three years. and spatial resolution, thereby providing the ability to recover certain ocean topography features or processes that are precluded with traditional altimeters. Such system would require a high gain, multi beam antenna, with the ability to steer the beams to track up to ten Mesoscale ocean eddies, analogous to atmospheric moveable reflection points over the ocean. storms, are thought to play a very significant role in the transport of momentum, heat, salt, nutrients, and other To demonstrate the feasibility of this measurement, GPS chemical properties of the ocean. At present, quantifying reflections were collected from airplane experiments over the role of mesoscale eddies in the ocean circulation and the ocean off the coast of California. The precision in the therefore climate variability cannot be done simply sea surface height obtained thus far is 5 cm over 1 sec, because their spatio-temporal structures are not resolved with 5 km spatial resolution, using two satellites. by the existing remote-sensing techniques. Current Additionally, an experiment with a fixed receiver at a generation altimeters have repeat cycles of at least ten lakeside indicated a precision of 2 cm, again using 2 days and cross-track separation of up to 300 Km. satellites. Such precision is suitable for eddy monitoring where sea level signals are usually 10 cm or more. Recent research efforts have begun using GPS signals scattered off the ocean and sensed by an air- or space- The R&D effort carried out at JPL during the past four borne receiver in a bistatic radar geometry, as a means of years has demonstrated that GPS altimetry is feasible doing altimetry and scatterometry. Upon impinging on from moving platforms, and its precision is useful for the ocean surface, the GPS signal is reflected primarily in scientific applications. The next steps involve the the specular (forward) direction, in an amount dependent development of GPS receivers which can track reflections on surface roughness and angle of incidence. An airborne and process the signals on board and, eventually, the or space-borne receiver, connected to a down-looking integration with suitable antenna systems which can antenna, could collect such scattered signals. One such capture the wealth of reflections available over the ocean receiver, and the 24 transmitters, form a multistatic radar at all times. system, capable of intercepting reflections from several areas of the ocean simultaneously. Because of the multi- static nature of the GPS observations, they will improve References our current capability of global sea surface measurements in two important ways: improved spatio - temporal C. Zuffada, T. Elfouhaily and S. Lowe: Deriving Near-Surface Wind Vector With Ocean Reflected GPS Signals: resolution and coverage. Simulations And Measurements, submitted for

Zuffada: Ocean Remote Sensing with GPS 65

publication in Remote Sensing of the Environment, March Lowe, S. T., C. Zuffada, Y. Chao, P. Kroger, J. L. LaBrecque, 2001. L. E. Young, Five-cm-Precision Aircraft Ocean Altimetry Using GPS Reflections, accepted for publication in A.K. Fung, C. Zuffada and C. Y. Hsieh, Incoherent Bistatic Geophys. Res. Lett., February 02. Scattering from the Sea Surface at L-Band, IEEE Transactions on Geoscience and Remote Sensing, Vol. 39, R.Treuhaft, S. Lowe, C. Zuffada, and Y. Chao: 2-Cm GPS no. 5, pp. 1006-1012, May 2001. Altimetry Over Crater Lake, Geophys. Res. Lett., 28, 23, p. 4343-4346, December 2001.

Journal of Global Positioning Systems (2002) Vol. 1, No. 1: 66-67

GPS For Ionospheric Sensing: Space and Ground Based

E. A. Essex CRC for Satellite Systems, La Trobe University, Vic 3086, Australia

Received: 5 May 2002 / Accepted: 16 June 2002

especially at low latitudes in the equatorial regions and in the high latitudes of both the Northern Polar regions and Bibliography Antarctica. Of particular interest is the reliability of WAAS during severe ionospheric disturbances. E. A. Essex, is a senior lecturer in the Department of Physics at La Trobe University. She obtained a PhD in Space Physics from the University of New England. Occultation Studies of the Ionosphere Since obtaining her doctorate, she has worked overseas at NASA originally developed the occultation (limb the University of the West Indies, and also at the Air sounding) technique for the study of planetary Force Geophysics Laboratory, Massachusetts. Currently atmospheres. With the advent of the Global Positioning she is the leader of the GPS Space Science project for the System (GPS) satellites accurate occultation Australian satellite FedSat. measurements of our own earth's atmosphere are now possible. As occultation occurs, the effects of the ionosphere on ray paths as they propagate from the GPS satellites to a receiver on board a low earth orbit (LEO) Space Weather Studies Using GPS satellite provide information on the ionosphere. These ray paths provide horizontal slices through the ionosphere. At the frequencies used by GPS, the ionosphere produces The information contained in these sets of signals is then the largest error in position location and timing. The extracted by an inversion technique known as magnitude of the total electron content (TEC) of the tomographic reconstruction to yield the vertical electron ionosphere is an important parameter in satellite density profiles of the ionosphere. measurements such as those obtained from GPS as it is directly related to timing and positioning determinations. As part of the activities of the Cooperative Research During extreme space weather effects such as magnetic Centre for Satellite Systems, Australia is planning the storms, the TEC of the ionosphere is often subject to launch of a scientific satellite, called FedSat, in large spatial and temporal variations over the globe celebration of the centenary of Federation. One of the (Essex et al, 2001(a)). payloads on board is a Blackjack GPS receiver, which will be used to undertake occultation studies of the With the occurrence of increased solar activity of the past studies of the ionosphere as well as navigation. Particular two years, around the peak of the sunspot cycle, interest emphasis will be directed toward the improvement of the in the possibility of severe space weather has increased. current models of the ionosphere over the vast expanses Many organisations around the world are now actively of the Southern Hemisphere where existing data is sparse. engaged in monitoring, modelling and predicting the This new, cost-effective technology based on the GPS space weather. One of the consequences of severe space constellation, is currently being trialled by space weather is often the disruption to satellite scientists using data bases from the German satellite communications and navigation. GPS navigation CHAMP and the Argentine satellite SAC-C. With a accuracy is particularly sensitive to changes in the constellation of LEOs equipped with on board GPS electron density profiles of the ionosphere. Dual receivers providing hundreds of occultations per day, frequency data from the constellation of GPS satellites near real time updates of the state of the ionosphere and ground stations are being used to study recent severe would be possible. This remote sensing technique offers magnetic storms and their effects on the ionosphere, not only a new data source for the upper atmosphere but

Essex: GPS For Ionospheric Sensing: Space and Ground Based 67 also may revolutionize the weather forecasting in the Kubik and N. Talbot, Canberra, Australia, 24-27July, lower atmosphere. 2001(a). CDROM. Essex, E.A., Birsa, R. Shilo, N.M., Thomas, R.M., Cervera, M.A. and Breed, A.M. Scintillation effects on Global References Positioning Signals under solar maximum conditions, Invited paper. Proceedings of the International Beacon Essex, E.A., Birsa, R. Shilo, N.M., Thomas, R.M., Cervera, Satellite Symposium. Ed. P. Doherty. Boston, USA, June M.A. and Breed, A.M. Global Positioning System signals 4-6, 2001(b). CD Rom under solar maximum conditions, SatNav 2001. Proceedings of the 5th International Symposium on Satellite Navigation Technology and Applications. Eds. K.