New Flexible Network-based RTK Service in Japan

New Flexible Network-based RTK Service in Japan I. Petrovski, S. Kawaguchi, M. Ishii and H. Torimoto, DX Antenna Co. Ltd., Tokyo, Japan K.Fujii, K.Ebine, Hitachi Ltd., Tokyo, Japan K. Sasano, Asahi National Broadcasting Co. Ltd., Tokyo, Japan M.Kondo, K.Shoji, Secretariat of High Precision RTK Committee (Tentative), NTT Communications Corp. H. Hada, K.Uehara, Y. Kawakita, J. Murai, WIDE Project, Japan T. Imakiire, Geographical Survey Institute, Tsukuba, Japan B. Townsend, Roberton Enterprises Ltd., Calgary, AB, Canada M.E. Cannon and G. Lachapelle, Dept. of Geomatics Engineering, University of Calgary, Calgary, AB, Canada BIOGRAPHY Ivan G. Petrovski is the Chief Researcher at the GPS Department of DX Antenna Co. Ltd. He holds MS and Ph.D. degrees in Aerospace Navigation from the Moscow Aviation University (MAI). Prior to joining DX Antenna, he was working as an Associate Professor at the MAI, and then as a Science and Technology Agency fellow with National Aerospace Laboratory, Japan Seiya Kawaguchi joined GPS department of DX Antenna in 1998. He holds a Master degree in Earth Science from National University of Kyusyu. Makoto Ishii is a Marketing Manager of the GPS department of DX Antenna. He was engaged in GPS business since 1987. Hideyuki Torimoto is a General Manager of the GPS department of DX Antenna. Before joining DX Antenna he had established Trimble Navigation Japan Ltd. and was working as an Executive Vice President of this company since 1986. Kenjirou Fujii holds MS from Waseda University. He is working at the Industrial Components and Equipment in Hitachi Ltd. Japan. He is a principal specialist in automatic control, robotics and GPS related system development. Kazuhide Ebine is a leading authority in the development of GPS related market and technology in Hitachi Ltd. He is engaged in the development of new areas of GPS related business applications. He is a Deputy Manager of the Sales Planning Dept. of Industrial Components and Equipment in Hitachi Ltd. Japan. Kouji Sasano works for Asahi Media Strategy Office. He had been engaged in developing fax broadcasting, data broadcasting and navigation systems. Masanobu Kondo is Senior Manager of the Value Development Office in Solution Business Division of NTT Communications Corporation. He is Technical Secretary of High Precision RTK Committee (Tentative). Kimiaki Shoji is a General Manager of Cross-Industry Sales and Marketing Dept. in Solution Business Division of NTT Communications Corporation. He is a chairman of High Precision RTK Committee (Tentative). Hisakazu Hada is an Instructor in the Digital Library Research Division at the Nara Institute of Science and Technology, Japan. He is a group leader of GPS related activity in the WIDE project. Keisuke Uehara is a Research Assistant in the Environmental Information College of the Keio University. He is a member of the WIDE project. Jun Murai is a Professor in the College of Environmental Information at the Keio University. He is the leader of the biggest Internet research and development consortium - WIDE project. He holds a Ph.D. in Computer Science from the Keio University. Dr. Murai is the President of the Japan Network Information Centre (JPNIC), a Member of Board of Trustee of Internet Society (ISOC), the Internet Corporation for Assigned Names and Numbers (ICANN), and Interlim Board of Directors. Tetsuro Imakiire is Senior Officer of Earthquake Investigation in the Geodetic Observation Center, Geographical Survey Institute (GSI). He is in charge of the management of GEONET(GPS Earth Observation NETwork), which is operated by GSI to monitor the crustal deformation of the Japanese islands. Bryan Townsend received his Masters of Science degree in 1993 from the Department of Geomatics Engineering at the University of Calgary. Since then he has worked in several areas of GPS including GPS surveying, GPS receiver design and wide area reference systems. Currently he is working in the area of Network RTK. Dr. Gerard Lachapelle is Professor and Head of the Department of Geomatics Engineering where he is

responsible for teaching and research related to positioning, navigation, and hydrography. He has been involved with GPS developments and applications since 1980. Dr. M.E. Cannon is a Professor in Geomatics Engineering at the University of Calgary where she conducts teaching and research related to GPS and integrated GPS/INS systems. She is a Past President of the Institute of Navigation. ABSTRACT This paper introduces a new approach to a Networkbased real-time kinematic (RTK) service, and presents the implementation and testing of such a system in Japan. The system implements a Virtual Reference Station algorithm, Internet based reference station network and TV sound multiplexed sub-carrier data link. The first implementation of the Network RTK system has been developed and tested in the Tokyo area. The network covers a region of approximately 100 km in radius. The service concept, components and test results for the proposed service are presented. INTRODUCTION This novel approach has advantages because it uses a combination of media for the data links between reference stations and the control station, as well as the control station and user. The approach uses Internet based reference stations for the reference network and a TV audio sub-carrier signal to transmit the RTK corrections data to the user. The Asahi TV broadcast system has been officially adopted by the Ministry of Post and Telecommunications for broadcasting DGPS and RTK corrections in Japan. For this test, six Ashtech Z-18 receivers were used for the reference stations, five of which are connected to the Internet through a dedicated server and one through an ISDN line to the Internet provider. This combination of different data links gives the proposed Network RTK service both scalability and expandability. There are two different Figure 2 Application of ASC data channel. approaches to the RTK service in Japan. One is a radiobroadcast service, which is scalable, but due to the bandwidth limitation, it is usually limited in terms of the amount of information transmitted. For example, it can supply the GPS user with either DGPS or RTK corrections. The other service is the Internet-based DGPS/RTK service, which at present can usually supply only a high latency data link (cellular phone or PHS) between the mobile user and the reference station. The proposed service combines advantages of both methods. It is scalable because it can serve an almost unlimited number of users in the coverage area, and it is expandable because the number and location of the Internet-based reference stations can be increased without disrupting service. These features give the service provider flexibility to expand the service based on market needs. The proposed service allows for a variety of GPS receivers and data-link systems to be used in the implementation. For the reference station network presented here, GPS receivers from at least three manufacturers have been tested on the user side. The implementation of the different equipment is possible due to the specifications of the data link based on the TV broadcast audio sub-carrier that allows the transmission of both DGPS and RTK correction data to the user in RTCM Version 2.2 format both for GPS and GLONASS. The system is currently implemented as a self-contained system in Tokyo. The service proposes an option that will allow the user to receive corrections from either a TV broadcast or directly from the control center through the Internet. This option gives the user more flexibility in case of a low latency data link to the Internet is used or if specific services are need, or alternatively if only DGPS corrections are required. In the near future, other areas around Japan will be covered. DEVELOPMENT OF COMPONENTS Figure 1 Internet Based Augmentation Network Concept In 1998, DX Antenna proposed an Internet Based Augmentation Network (IBAN) concept that was based

position. The idea and algorithm of a correction grid calculation was similar to satellite based Wide Area Augmentation System (WAAS), lacking however its integrity and range service. The advantage of the Internet-based augmentation service is that it can easily be expanded and easily used without any special equipment. Figure 3 Wireless mobile data link latencies mapped over user track. upon the Internet technology developed by the WIDE project (see Figure 1). WIDE is the biggest Internet research group in Japan that includes over one hundred participants from different companies and universities. A special research group in WIDE had developed Internet-based Differential GPS (DGPS) and Real Time Kinematic (RTK) reference stations (see Hada at al., (2000)). This Internet-based DGPS/RTK service is designed to provide GPS users with differential code and carrier phase corrections through the Internet. The user can select an Internet-based reference station that is located near their position. The nearest position could be different in terms of geographical location and the Internet. Specially developed software called propagation agent finds the optimal server for the particular user. The Internet-based DGPS service allows the creation of a low-cost infrastructure that could be used worldwide. Bi-directional communication between a server and a GPS user allows the transfer of only required information through the optimal channel. The user can select a subset of the correction information that is needed for their particular purposes, and the propagation agent can select the best server for this user. The IBAN supplies the user with differential corrections, calculated based on the information from a number of servers, which are located near the user Figure 4 ASC coverage for two different antenna types. After SA was turned off, the necessity for a DGPS service becomes less obvious in Japan. On the other hand, different applications require more precise (centimeter level) accuracies. Therefore, development of an RTK service becomes an essential task. At present, the Internet-based reference stations are not suitable for RTK due to the large latencies in mobile data links. Although the Internet itself can provide reliable a data stream using TCP/IP protocol with latencies less than one second, wireless mobile communication devices have low reliability and demonstrate latencies from 3 to 8 s for typical tests. Figure 3 shows latencies over a 40 minute car test Figure 5 VRS Concept through an urban area. There are also frequent cutoffs that result in latencies up to 15-20 s due to reacquisition of the data link. Therefore, one can conclude that Internet-based reference stations for an RTK service are very convenient when using a wired data link for static user, and not reliable for a mobile user having a wireless data link. Approximately at the same time Asahi TV together with DX Antenna began to transmit differential and RTK GPS/GLONASS corrections encoded into TV audio sub-channel signal (ASC). This service was intended for mobile users such as car navigation systems. Figure 2 depicts the concept and required user equipment for the ASC data channel. Apart from a GPS receiver, the user is required to accommodate an ASC receiver. Besides RTK and DGPS/DGLONASS corrections, the

ASC receiver provides an extra data channel for weather and traffic information. The main disadvantage of this service in terms of RTK was that RTK is generally used at distances of 10-15 km from the reference station, while the ASC data broadcast is available from 40 to 100 km, depending on the antenna type. Figure 4 demonstrates the area coverage for two different types of antennas. An ASC-based RTK service also has potential to become nationwide, because of the existing infrastructure of the Asahi TV station network and the satellite TV stations. RTK NETWORK To combine flexibility, expandability and availability of the Internet-based reference stations with scalability, Figure 8 Demo RTK Network in Tokyo allows modeling of the orbital, ionospheric and tropospheric errors, which are decorrelated with distance. A correction grid, which is calculated from data from all network reference stations, is transferred to the user as data from a new reference station that exists in the user s vicinity. Therefore, data from the entire network is encapsulated into one reference station. Figure 6 Number of required reference stations low latencies and high capacity of the ASC channel, the proposed RTK Network system had been developed. The proposed system uses a Virtual Reference Station (VRS) concept, which is rather well known from number of papers (see Figure 5). An RTK network There are a large number of GPS users in Japan who are interested in high accuracy GPS positioning and therefore use the RTK method. These users belong to the user segment which includes the survey, navigation, and construction fields. Despite the high density of publicly available reference stations (the largest GPS network consists of about 1000 reference station and belongs to the Geographical Survey Institute), some stations do not provide real time data. Most users still Figure 7 RTK VRS Demo System

Figure 11 ASC data channel have to use their own reference stations due to the fact that atmospheric conditions over Japan cause rapid changes with ionospheric and tropospheric errors over distance. These users will benefit from the use of a network of permanent GPS reference stations. This network will provide users with more accurate and more robust solutions than the single reference station solution. Also, it relieves them of the burden of supplying their own reference station and data link. The main purpose of the RTK network is to suppress ionospheric influence on the GPS results. The characteristics and conditions of ionosphere define the configuration of the network itself. Basically, the VRS concept allows a reduction in the number of reference stations required to cover any given area (see Figure 6). At the same time, integrity and reliability of service will increase in contrast with a single reference station with the same or better accuracy. As a result of these demands and existing technological components, it was decided to develop an RTK Network infrastructure that allows seamless RTK over all Japan combining Internet-based reference stations with an ASC data link. VRS RTK NETWORK DEMO SYSTEM To evaluate the proposed RTK Network system, a demo system was established in the Tokyo area. A layout of the demo system is depicted on Figure 7. There are six reference stations that are available through the Internet. The reference stations are connected to the control station, which includes a GPS reference station receiver, an ASC receiver and a server computer. The control center calculates correction data for the users in the area. As a core engine for the control center the DeciPos/Multi-Reference RTK method is used, which was proposed and developed by Roberton Enterprises (see Townsend et al (1999)) and the University of Calgary (Racket et al (1998)). Figure 9 Primary RS, and ASC receiver on top of monitor station There are different types of VRS algorithm implementations. There are two main approaches, one which adds the correction grid to the existing reference station data, and another one which creates the VRS raw data itself. The algorithms also differ in terms of communication Figure 10. Control Center antenna installation

reference stations varies from 50 to 100 km, and the broadcast distance is up to 70 km. Figure 8 shows a map of Tokyo with reference station locations depicted by red triangles and the correction grid depicted by blue circles. An example of the prime VRS is depicted by a dark blue triangle. Six secondary reference stations belong to the WIDE project, GSI and the Control Research Laboratory (CRL). All reference stations accommodate geodetic quality L1/L2 GPS receivers. The Control Station with the primary reference station was located at the Open Network Ltd. office. Besides the main server, the Control Center accommodates a backup server and another GPS receiver as a monitor station. An ASC receiver feeds the monitor station with broadcast corrections. Figure 9 demonstrates JPS control station receiver and ASC receiver on top of monitor station housing. Figure 10 shows antenna installation on top of Open Network building. Calculated VRS data are then transferred through a 64 Kbps dedicated line to Asahi TV facilities (see Figure 11). These data are encoded into TV audio signal by a DQPSK modulator, combined with a video signal in an audio transmitter, which is transferred to an Asahi TV tower and then is broadcast to a user. There also a number of satellite TV stations that cover areas which are shadowed from the TV tower by mountainous topography. Figure 12 RTK test results for a rover at the different positions inside the Network. link between the control center and the user. There are one-directional and bi-directional communication links. In the second case, the user gives its approximate position and allows the control center to calculate the VRS or corrections from the VRS in the vicinity of the user. In the first case, theuser calculates VRS data based on the grid data. The algorithm used lies between these two methods and utilizes two VRSs. The first VRS is calculated at the control center and is placed in the middle of a predefined area, while the second one is calculated by the user. This allows for easy expansion of the control center service for a larger area, which requires two or more primary VRSs. This algorithm also incorporates options for one- and bi-directional communication. A bi-directional communication algorithm, generally speaking, can put some extra constraints on the number of users. In Japan, the potential number of users could reach a few thousand for one service area, which requires special equipment and extra time to connect an unknown number of users to the control center server at the same instant. This situation can cause increased latencies, which is a very critical issue for RTK methods. Figure 13 Robotic device installation for kinematic RTK test. The user is equipped with a standard off-the-shelf single or dual frequency receiver capable of performing RTK positioning, a personal computer and an ASC receiver. The required equipment should implement standard RTCM V2.1/2.2 corrections format message Type 1, Type 18/19, Type 20/21, Type 3, and if required, Type 59 for area correction parameters. The baud rate is between 2400-4800 bps. Basic requirements and characteristics of the RTK network were based on the Geographical Survey Institute (GSI) specifications. Distance between

Figure 15 Tractor with automatic control system, which use VRS RTK and ASC. demonstration included static, kinematic tests and automatic control of a tractor, using VRS data. To ensure the system accuracy in the RTK mode a special robotic device has been implemented (see Figure 13), which allows a calibrated GPS antenna movement along a three meter rail from the surveyed point. Figure 14 shows the kinematic test results. Figure 15 shows the tractor moving in automatic mode without a driver. The onboard navigation system uses RTK GPS data to control tractor movement. The tractor has NovAtel GPS receiver, an ASC receiver and an INS installed. FUTURE DEVELOPMENT OF RTK NETWORK Figure 14 RTK measurements of the antenna movement along a rail of the robotic device. DEMO RTK SYSTEM TEST RESULTS. INFRASTRUCTURE Figure 16 shows the proposed concept for RTK network nationwide coverage. Every one of the nine segments will accommodate one control center, which calculates A number of tests conducted at different locations over the network show that an RMS of about 3 cm is achievable during rather harsh ionospheric conditions, that cause error decorrelations of around 14 ppm. Figure 12 demonstrates three scatter plots of the rover location relative to the reference stations. In fixed ambiguity mode, the accuracy is uniformly better than 5 cm. In float mode, given that a quality L1 receiver is used, the accuracy is uniformly 10 cm, after a specified initial convergence time. For the test at Nara, the shortest baseline was 28.7 km. The tests also show the advantage of this multi-reference RTK Network is nearly uniform in accuracy, regardless of the proximity of any one reference station. The main demonstration of the RTK Network was conducted at the Hitachi Ltd. territory and attracted more than 300 specialists from different companies, universities and research organizations. The Figure 16 Proposed VRS nationwide coverage.

a number of primary VRSs. The size of each segment depends on the ASC bandwidth, which should accommodate corrections grid for all area. The segments are overlapped in order to provide seamless integration. There is an extra channel for a bidirectional communication mode through the Internet apart from the nine ASC oriented control centers. Each segment is similar and is presented by Figure 17. There are also a number of towers, because the broadcast area of one TV tower is less than one segment. The Internet is included as an option for bi-directional communication data link through mobile telephone. The existing software and infrastructure, with relatively minor modifications, will be able to accommodate more GPS satellites, additional GPS frequencies, and a new system like GALILEO. The advantages will be increased robustness, shorter time to integer ambiguity resolution and higher accuracy. There will be more satellite measurement data to transmit but thanks to the effectiveness of the approach, the effect on the correction data will be minimal. SUMMARY An RTK network infrastructure has been developed and successfully tested in Tokyo area. The developed system is using virtual reference station concept, the Internet based reference station network and audio subcarrier data channel. The system is intended for nationwide coverage. REFERENCES 1. H. HADA, H. SUNAHARA, K. UEHARA, J.MURAI, I. PETROVSKI, H. TORIMOTO, S. KAWAGUCHI (2000). DGPS and RTK Positioning Using the Internet. GPS Solutions Vol.4, N1, John Wiley & Sons, Inc. 2. FORTES, L.P., G. LACHAPELLE, M.E.CANNON, G. MARCEAU, S. RYAN, S. WEE and J. RAQUET (1999) Testing of a Multi- Reference GPS Station Network for Precise 3D Positioninng in the St.Lawrence Seaway. Proceedings of GPS99 (Session A4, Nashville, 14-17 September), The Institute of Navigation, Alexandria, VA. 3. TOWNSEND, B., G. LACHAPELLE, L. FORTES, T.E. MELGARD, T. NØRBECH, and J. RAQUET (1999). New Concepts for a Carrier Phase-Based GPS Positioning System Using a National Reference Station Network. Proceedings of National Technical Meeting, The Institute of Navigation (January 25-27, San Diego, CA), 319-326. 4. RAQUET, J. G. LACHAPELLE, and L. FORTES (1998) Use of a Covariance Analysis Technique for Predicting Performance of Regional Area Differential Code and Carrier-Phase Networks. Proceedings of GPS98 (Session A5 (Nashville, 15-18 September), The Institute of Navigation, 1345-54. Figure 17 Control segment layout