FIELD TEST OF THE GPS+GLONASS RTK AT THE CAGLIARI PERMANENT STATION VIA INTERNET Giannina Sanna, Giuseppina Vacca University of Cagliari Department of Structural Engineering P.zza D Armi 913 Cagliari, Italy (topoca, vaccag)@unica.it tel. +3977 fax +3977 KEY WORDS: Surveying, GLONASS, GPS, Internet, Real-, RTK. ABSTRACT: In the frame of the EUREF-IP project, the Topography Section of the University of Cagliari has set up a permanent station, which since July sends differential corrections via Internet for both DGPS and RTK positioning. The permanent station, identified by the name CAGZ, is part of the IGLOS network and, since October 3, of the EPN network. The Javad-Topcon permanent station receiver acquires signals from both the GPS and GLONASS constellations. In this job the results of various tests are introduced, in which GLONASS observations have been added to the GPS ones to study the improvement obtained by the RTK technique. 1. INTRODUCTION Today various international studies are addressed to the improvement of real time kinematic positioning. RTK requires a GPS reference station sending raw data to a rover receiver simultaneously tracking the same satellite signals, to form double difference observations and obtain centimeter level positioning. Some requirements limit the occurrences under which precise positioning is obtained, especially in motion for surveying or for precise navigation applications on land. These requirements concern: the range and reliability of the broadcasting systems of data transmission; the distance between master and rover; the OTF algorithms of initialization; the number of satellite in view. In this work two aspects are involved, that concern the data transmission and the satellite constellation. These aspects are tied to two improvements brought recently, namely the use of Internet as a medium for transmitting GNSS data and the increase of the number of available satellites due to the plan, from the Russians, to restore the GLONASS system. Regarding the first item, the EUREF (EUropean REference Frame) started in June a project named EUREF- IP, with the purpose of developing a stable and robust infrastructure for broadcasting differential corrections via Internet. Our research group has joined the EUREF-IP project carrying out several tests regarding protocols of transmission (simple tcp/ip or http) under various transmitting medium (wireless lan, GSM, GPRS). Currently we have a range from 9. kbit/s (with a GSM modem) to 7. kbit/s (with a GPRS connection), up to a potential rate of Mbit/s for the UMTS network (not yet completely operative in Italy). Our study has confirmed the improvement brought by the protocol developed by EUREF, but it has put in evidence as the GPRS, valid in the first stage of its implementation, is now dramatically influenced by network overload conditions. The second aspect on which currently the study is focused concerns satellite visibility, which can be dramatically reduced in urban areas or forest environments. The loss of the GPS signal brings to an intermittent distribution of the positions fixed with precision, both for the lack of the minimum number of satellites and for extend of the reinitialization phase due to continuous changes in constellation. This shortcoming will be overcome with the launch of the GALILEO system but, until then, GLONASS, although not yet complete, provides extra satellites increasing the in-view constellation. In this job the results of various tests are introduced, in which GLONASS observations have been added to the GPS ones to study the improvement obtained by the RTK technique.. THE GLONASS SYSTEM The GLONASS (GLObal NAvigation Satellite System), managed by the Russian Space Forces, when fully deployed will be composed of satellites in three orbital planes. The GLONASS system has two types of navigational signals: standard accuracy signal and high accuracy navigation signal. The standard accuracy signal with clock rate.11 MHz is designed for civil users. The high accuracy code with clock.11 MHz is modulated
by special code and its unauthorized use is not recommended. Table compares GLONASS and GPS spatial segments. Parameter GLONASS GPS Number of satellites Number of orbital planes 3 Orbital altitude 191 km 183 km Inclination.8 Orbital period 11 h 1 min 1 h Ground track repeatability every eighth sidereal day every sidereal day Frequency band L1 1-1 MHz L 1 1 MHz Tab. : GLONASS vs GPS System L1 17. MHz L 17. MHz The GLONASS broadcast ephemeris describes a position of transmitting antenna phase center of given satellite in the PZ-9 Earth Fixed Reference Frame, slightly different from WGS8. At the moment ( December ) the system GLONASS constellation is as composed of 11 satellites. A launch of 3 new satellites is scheduled on December. One of these will be a GLONASS-M (with a seven-year service life) and the other two will be the older model satellite (with a three-year service life). This will bring the total number of GLONASS satellites in orbit to 1. The Russian Aerospace Agency has the approval of the Russian Government to continue a long-term plan for the period -11, during which plans to reconstitute the full constellation. 3. THE CAGLIARI PERMANENT STATION Fig. 1: CAGZ location The permanent station of University of Cagliari (Italy), is situated into the Astronomic Observatory of Cagliari (fig.1 and ). The station consists of a GPS+GLONASS Javad-Topcon Legacy/E receiver with external frequency source (cesium) and a Regant- choke-ring antenna. The permanent station, identified as CAGZ, is part of the IGLOS network and, since October 3, of the EPN network. On the web page http://www.epncb.oma.be/_trackingnetwork/info/cagz.html are some information about the CAGZ Permanent Station as station logs, data availability, data quality etc. The receiver firmware provides the RTCM messages for code-differential and RTK corrections and since July are broadcasted differential corrections over the Internet for DGPS and RTK positioning. The Internet servers run on a PC that is connected to the receiver (via serial cable) and to the Internet. The transmission of the differential corrections, in the RTCM format, is performed by two server applications: DGGI (Differential GPS and GLONASS via Internet), developed by the Cagliari research group and NTRIP infrastructure (Networked Transmission of RTCM over IP) Fig. : The Permanent Station CAGZ developed by the EUREF. The DGGI transmits DGPS corrections (RTCM 1 and 31 messages) every 3 seconds while the station information (messages 3 and 1 with the station name and coordinates) are sent every 3 seconds. Further information about DGGI Server are in the web page http://laser1.ca.astro.it/gps/rtcm.html. The service, which is online since July, is free of charge. Fig. 3. Scheme of the Permanent Station with the RTCM Servers The NTRIP Server, supplied by the EUREF-IP and based on the HTTP/1.1 protocol, is adapted to support the streaming of GNSS data. The system is implemented in three applications, named NtripServer, NtripClient and
NtripCaster. The former two are technically HTTP clients, while the latter is the true HTTP server. The NtripServer software runs on the permanent station PC, and as a client connects to the NtripCaster on the port 8 or 11, then sends to it the differential corrections as the receiver produces them. The NtripCaster is a HTTP server and it is enabled to send data to many users (even thousands) at the same time. The NtripClient, or GNSS Internet Radio, runs on the client computer (be it a common PC, a lap-top, or a handheld system with Windows CE), which is connected to the rover GPS receiver. Upon starting, the user selects the station whose RTCM correction will be used; then, the software connects to the chosen station and starts downloading the corrections, which are sent to the receiver through the serial port. The user, that wants to use the differential corrections by CAGZ, has to connect with NTRIPClient to NTRIPCaster, which is installed in Frankfurt, so he can choose the CAGZ Permanent Station. Figure 3 shows the behaviour of CAGZ in the year. The high rate of outages, specially for a few months, depends on a bad working of the receiver. Monthly outages () 8 Jan. Feb. Rate % March April May June July Aug. Sept. Oct. Nov. Dec. Fig. : CAGZ outages. STATIC TEST The equipment was based on a Javad-Topcon GPS+GLONASS L1L geodetic receiver with a LegAnt antenna connected to a notebook PC, in turn connected to the Internet by a GPRS modem (Nokia D11). The receiver applied phase differential corrections from the CAGZ reference station. All measurements were performed with only one GPS+GLONASS receiver so, in order to avoid differences in GPS satellite configuration, tests were scheduled at almost the same time in two different days (1/ May ). The RTK tests using only GPS constellation are pointed out as RTK-GPS (Real Kinematic - GPS), while those using both GPS and GLONASS constellations are pointed out as RTK-GG (Real Kinematic GPS+GLONASS). Three tests RTK-GPS and three RTK-GG were carried out on known point placed at different distances from the permanent station. The point called SHORT was a few meters from CAGZ, while the point called LONG was about 1 km far from CAGZ. The time span was 1 h and the observation rate was 1 HZ with 1 degrees elevation mask. Fig. : Equipment Fig. : Static test
For every test the following parameters were investigated: - rate of fixed solution - number of satellites in view - HDOP parameter - time to correctly fix ambiguities To establish if the ambiguities were correctly fixed, a specific receiver routine, comparing the known coordinates with those computed in RTK mode, was used. If the difference on each component is under cm, the estimated coordinates are considered correct. The values and the time to assess them, together with the raw data are recorded. Then the RTK engine is reset. Table 3 shows the mean rates of solution type achieved in the tests: - Single position in which the correction didn t arrive - Fixed solution in which integer ambiguities were fixed - Float solution in which ambiguities were solved for as real number. Table shows the mean number of satellites tracked, the HDOP parameter and the time to correctly fix ambiguities. The results show that the number of fixed position, in RTK-GG, increases about 1% in the short baseline and 17% in the long baseline, while the time to fix ambiguities is halved for both baseline with respect to RTK-GPS. Point SHORT LONG Technique Solution RTK-GPS RTK-GG RTK-GPS RTK-GG Single Point 38 % 17 % 18 % 17 % Fixed Solution 3 % % 1 % 38 % Float Solution 3 % 1 % 1 % % Tab. 3: Mean of the rates of solution Static test KINEMATIC TEST Point SHORT LONG Technique Mean value RTK-GPS RTK-GG RTK-GPS RTK-GG N. satellites 7.7 8.8 7. 1. HDOP 1..8.8.8 to correctly fix 7.7 s 3. s.1 s 3.3 s Tab. : Mean of number of satellites, HDOP and time to correctly fix Static test The kinematic tests were performed in an urban area, 1 km far from the reference station, characterized by buildings, trees and open fields. The equipment was set up on a car driving at about km/h. The path, about km long, was covered performing times in sequence RTK-GPS and RTK-GG tests. Table shows the mean rates of solution types both in RTK-GPS and RTK-GG. Table shows the mean number of satellites and the mean HDOP. Figure 9 and 1 show the trajectory of the vehicle in both RTK-GPS and RTK-GG mode. The dark colour indicates fixed solutions, the grey colour indicates float solutions, the blank colour indicates single point solutions. Figure 11 a, b, c and 1 a, b, c show changes in the number of the satellites, HDOP parameter and solution type during experiments. Figure 13 shows the satellites elevations on 9 May when the kinematic test were performed. The results achieved in RTK-GG tests point out a greater rate of fixed positions, about 7%, with respect to RTK- GPS, regardless of the greater amount of single point solutions in RTK-GG. Figure 13 shows that during all kinematic tests, the GLONASS constellation was of satellites with a maximum elevation lower than 3 degrees. Point Technique RTK-GPS RTK-GG Solution Single Point 3 % % Fixed Solution 13 % % Float Solution % % Tab. : Rates of solutions Kinematic Test
Point Technique RTK-GPS RTK-GG Mean Value N. satellite.3 7.3 HDOP 1.3 1.3 Tab. : Mean of n. of satellites, HDOP and RMS - Kinematic Test Fig. 7: Trajectory by RTK-GPS Fig. 8: Trajectory by RTK-GG N. Satellites RTK-GPS a N. Satellites RTK-GG a N. satellites 1 1 8 13 13 1 13 1 171 18 183 19 133 11 113 13 13 1 13 13 1713 183 18 19 113 111 111 N. svs 1 1 8 118 118 11 1173 1188 118 1193 19 18 113 11 18 133 11 1 13 17 18 178 181 18 193 138 13 1313 131 139 1333 1 HDOP RTK-GPS b 1 HDOP RTK-GG b HDOP 1 HDOP 1 13 13 1 13 1 171 18 183 19 133 11 113 13 13 1 13 13 1713 183 18 19 113 111 111 118 118 11 1173 1188 118 1193 19 18 113 11 18 133 11 1 13 17 18 178 181 18 193 138 13 1313 131 139 1333 Solutions RTK-GPS c Solution RTK-GG c Solutions 3 1 13 13 1 13 1 171 18 183 19 133 11 113 13 13 1 13 13 1713 183 18 19 113 111 111 Solutions 3 1 118 118 11 1173 1188 118 1193 19 18 113 11 18 133 11 1 13 17 18 178 181 18 193 138 13 1313 131 139 1333 Fig. 9: RTK-GPS Kinematic Test Fig. 1: RTK-GG Kinematic Test
Fig. 11: Satellites elevation on 9//. REMARK In order to assess the impact of combined GPS/GLONASS constellation as compared to GPS only, different RTK tests were performed. A comparison between the rate of the correctly fixed phase ambiguities and the time to correctly fixed ambiguities were investigated. Good results are obtained in the RTK-GG static test with an increase of the rate of the correctly fixed ambiguities (up to 17 %) and with a decrease of the time of correctly fixed ambiguities (up to %). Less impressive are the results in the RTK-GG kinematic tests. During the tests, GLONASS satellites were in view at a low elevation but, because of the presence of buildings and trees along the path, they were tracked only occasionally. This instability brought to a continuous change in constellation protracting the time to ambiguity resolution. In order to assess if a significant improvement could be obtained in precise navigation, more tests will be carried out with GLONASS satellites at the highest elevation, regardless of their number. REFERENCES [1] Pala A. et al.. Real Mapping with DGPS-Enable Navigation Equipment. In Proceedings of XX Congress ISPRS - 1-3 July Istanbul, Turkey [ISSN 18-1777] [] Pala A. et al., 3, L impiego di sistemi integrati GPS-PC palmari per il posizionamento di precisione in tempo reale e l acquisizione di dati spaziali, In Proceedings of 7 a Conferenza Nazionale ASITA, Verona 3, Vol. II: 139-1 [3] Weber, G., Gebhard, H., Dettmering, D., 3, Networked transport of RTCM via Internet Protocol, International Union of Geophysics and Geodesy General Assembly, Sapporo, July 3 [] Falchi E. et al.,, Internet-based DGPS service from the Cagliari permanent station. GIS application with LapTop and Pocket PC, Proceedings of the Seminar/Workshop Real GNSS Trieste in Reports on Geodesy Warsaw University of Technology N. 3 (3), [] In Su Lee et al.. The kinematic positioning of vehicle with Real differential GPS/GLONASS and Real- Kinematic GPS/GLONASS. In Proceedings of Integrated System for Spatial Data Production, Custodian and Decision Support ISPRS Commission II, Symposium China August -3, [] Weber G., 1, EUREF and Real Products Proceedings of EUREF Symposium, Dubrovnik, Croatia, May 1 [7] Hada H., et al., 1999, New Differential and RTK corrections service for Mobile Users, Based on the Internet, In Proceedings of ION GPS 99, Nashville, TN, 1-17 September 1999, 19-7 [8] Van Diggelen F. 1997. GPS and GPS+GLONASS RTK. In Proceedings of ION GPS 97, Kansas City, MO.