Examination of Site Suitability for GNSS Measurements

Transcription

1 TS 2 Geodetic Networks, Data Quality Control, Testing and Calibration Examination of Site Suitability for GNSS Measurements Danijel Šugar 1, Petar Sučić 1, Željko Bačić 1 1 Faculty of Geodesy, University of Zagreb, Kačićeva 26, Zagreb, Croatia, dsugar@geof.hr, psucic@geof.hr, zbacic@geof.hr Abstract. In order to make high-quality GNSS observations leading to reliable coordinates, especially on the stations that are intended to be used frequently or permanently, it is necessary that the location on which the observations will be taken fulfils certain criteria encompassing primarily the horizon without any obstacles, but also the absence of noise generated by the sources of electromagnetic radiation. The fulfilment of these criteria has been examined from the data taken during 24-hours long session of GNSS observations on the station located atop the tower of the Faculty of Mining, Geology and Petroleum Engineering (RGN) below which the antennas and the base station of mobile telecommunication provider are located. The observation data processing done with the TEQC software provided the information about Signal-to-Noise Ratio (SNR) and multipath both at L1 and L2 frequencies for signals of GPS and GLONASS satellites. Similarly, the observation data from the nearby CROPOS CORS ZAGR were processed and the comparison of results was presented. Based on GNSS observations taken on the RGN station along with the observation from CROPOS CORS stations KARL, SISA, ZABO and ZAGR, the baselines were processed, subsequently the network was adjusted and the coordinates accuracy estimation was given. Additionally, the apparent horizon around the RGN station was observed with the total station revealing that the sky is without obstacles above 10 elevation mask. Keywords: GNSS observations, horizon, multipath, Signal-to-Noise Ratio (SNR), TEQC. 1. Introduction The station sites that are intended to be frequently or permanently occupied are required to meet certain criteria for the purpose of capturing high-quality GNSS observations resulting in reliable coordinates, especially related to the horizon being completely free of obstacles, but also to the absence of noise generated by the sources of electromagnetic radiation. The absence of obstacles (horizon cleanliness) around a GNSS station is usually determined by simple terrain reconnaissance, but it can be determined more reliably by observing the zenith distances, i.e. the elevation angles of objects on apparent horizon. However, in order to detect the noise caused by the sources of electromagnetic radiation, it is necessary to carry out GNSS observations in a longer time window and 255

2 SIG 2016 International Symposium on Engineering Geodesy, May 2016, Varaždin, Croatia subsequently process the data using appropriate software tools and interpret the results. Such measurements were performed atop the tower of the Faculty of Mining, Geology and Petroleum Engineering (RGN) of the University in Zagreb, with the antennas and the base station of mobile telecommunication provider located below the tower [Figure 1.1]. Figure 1.1 Tower of RGN (GNSS station is on the top and monumented with metal pillar) Except the antennas shown on figure 1.1, the antennas and the base station owned by another mobile telecommunication provider are located on the roof of the RGN building, approximately 40 m apart from the tower. Fortunately, the latter are not oriented towards the tower, so the GNSS station is not located within the directed beam of radiation. The depicted circumstances with the sources of electromagnetic radiation and the potential noise effect on the quality of GNSS observations have motivated the research work and the preparation of this paper as well. 2. Multipath and Signal to Noise Ratio Multipath is mainly caused by reflecting surfaces near the receiver, secondary effects are reflections at the satellite during signal transmission. There is no general model of multipath effect because of the time- and locationdependent geometric situation. The influence of the multipath, however, can be estimated by using a linear combination of code and carrier phase measurements on frequency L1 and L2. The principle is based on the fact that the troposphere, clock error, and relativistic effects influence code and carrier phase by the same amount. This is not true for the ionospheric refraction and multipath, which are frequency dependent. Taking ionosphere-free code ranges and carrier phases, and forming corresponding differences, all mentioned effects except for multipath are canceled [Hofmann-Wellenhof et al. 2008]. Multipath propagation affects both code and carrier measurements. The effect on P-code observations is two order of magnitude larger than on carrier phase observations, and can reach decimeter to meters [Seeber 2003]. It is assumed that the phase multipath on carriers L1 and 256 Figure 1.2 GNSS receiver on RGN station (metal pillar) during 24-hours long session

3 TS 2 Geodetic Networks, Data Quality Control, Testing and Calibration L2 (mp1 and mp2) are small in comparison to MP1 and MP2 pseudorange multipath on carriers L1 and L2. Pseudorange multipath MP1 and MP2 can be estimated by equations [Seepersad & Bisnath 2015]: 1 = 1 1 2, (1) 2 = 2 1 2, (2) where P1 and P2 are pseudoranges on carriers f (L1) and f (L2), L1 and L2 are phase measurements on carriers f (L1) and f (L2), respectively. Coefficient α is derived according to [Seepersad & Bisnath 2015, URL 1]: =( ( 1)/ ( 2) ) = (3) A complete derivation of expressions (1) and (2) may be found in e.g. Zrinjski, As was shown in Seepersad & Bisnath 2015, at lower elevations angles there is higher multipath and as the observed elevation of the satellites increases, the multipath decreases. The occurrence of multipath at the GNSS antenna adversely affects GNSS initializations and solutions: during initialization by the OTF (On- The-Fly) method, it is difficult to detect multipath, which may cause initialization to be lengthy or unsuccessful [Trimble Navigation Ltd 2013]. The Signal to Noise Ratio (SNR or S/N) describes the performance of GNSS receiver functional block by relating the signal power P to the noise power N according to expression [Hofmann-Wellenhof et al. 2008, Seeber 2003]: =10. (4) The carrier-to-noise power density ratio C/N 0 is a bandwidth-independent index number that relates the (carrier) power to noise per 1 Hz bandwidth: =10. (5) Whereas S/N is generally used in conjunction with signal at baseband after dispreading operations, C/N 0 is more commonly used to quantify the signal power P r of received signal [Bradke 2009]. Typical values range from roughly 30 or 40 dbhz at low elevations to 50 to 55 dbhz at elevations above 60. This quantity is used for comparison of signal strength between channels and between satellites, and to assess interference. It can also be used to map the multipath environment around antenna, to estimate time-varying multipath parameters, and may be a way to remove multipath errors from phase data [Estey & Wier 2014]. An overview of Signal To Noise Ratio and carrier-to-noise is given in e.g. Joseph GNSS measurements on RGN station The observations atop the tower of RGN were made using the GNSS receiver Trimble R8 (model 2) with the integrated antenna R8 GNSS/SPS88x and the controller Trimble (Trimble Survey Controller). Receiver parameters were set as follows: data logging interval 10 seconds, elevation mask 0, recording of L2C, L5 and GLONASS satellites data. The data logging started on 22 nd February 2016 at 07:16 GPST (GPS Time) and ended on 23 th February 2016 at 07:34 GPST which means that the session lasted for more than 24 hours. The observations were made on the geodetic station monumented with 1.30 m high metal pillar 257

4 SIG 2016 International Symposium on Engineering Geodesy, May 2016, Varaždin, Croatia (tube). On the top of the metal pillar, there is an aluminum disk with the threaded screw in its centre for setting up the geodetic instruments [Figure 1.2]. A tribrach was mounted on the screw and used for levelling the GNSS receiver with integrated antenna. The antenna height was measured vertically from upper plane of the aluminum disk up to four positions evenly spaced along the center of bumper. As it will be explained later in this paper, the station coordinates were determined by means of static measurements. Horizontally, coordinates are related to the center of the screw, vertically to the upper plane of the aluminum disk. The power was supplied from external battery [Figure 1.2]. For the purpose of data processing, the comparisons of the results with the nearby CROPOS CORS ZAGR, as well as the baselines processing, network adjustment, and the determination of RGN station coordinates, GNSS observations for CORS stations KARL, SISA and ZABO were downloaded from CROPOS GNSS Reference Station WEB Server [URL 2] in RINEX 2.11 format (more details about this format may be found on e.g. URL 3.). The observations recorded with data logging interval of 10 seconds covered the time window from 22 nd February 2016 at 07:30 GPST up to 23 rd February 2016 at 07:30 GPST. 4. Data processing in TEQC TEQC is a comprehensive toolkit for solving many problems in the preprocessing of GNSS data. The most common data format used in TEQC is RINEX (OBS, NAV, MET files) being currently supported formats up to the version 2.11 [Estey & Lou 2014]. The usage of proper options in TEQC provided the reduction of original RINEX file with the observations gathered on the RGN station to 24-hours long time window matching the same time span used for the data on CROPOS CORS ZAGR. The time spans are given in GPST that is currently (March 2016) ahead from UTC by 17 seconds (leap seconds). In RINEX file (RGN station), there are 13 types of observations contained: C1, C2, C5, P2 (pseudoranges in unit m), D1, D2, D5 (Doppler frequency shift in unit Hz), L1, L2, L5 (carrier phase in unit full cycles), S1, S2 and S5 (Signal-to Noise Ration for phase observations in unit dbhz). It is specified in the User Guide for GNSS receiver Trimble NetR5 installed on CROPOS CORS stations that typical SNR of a satellite at 30 elevation is between 47 and 50 dbhz; the quality of a GPS position is degraded if the SNR of one or more satellites in the constellation falls below 39 [URL 3]. The header of RINEX file with data from RGN station is shown on figure 4.1. GNSS data in RINEX 2.11 format observed at CROPOS CORS include only 8 types of observation, so for the purpose of further data processing, analysis and comparison, the observables C2, C5, D5, L5 and S5 contained in RINEX file at station RGN were removed. 258

5 TS 2 Geodetic Networks, Data Quality Control, Testing and Calibration Figure 4.1 Header of original RINEX file with the observations gathered on the RGN station Additionally, as it is reported in figure 4.1, the model designation of GNSS receiver R8 Model 2 along with the designation of GNSS antenna TRM are not in accordance with IGS (International GNSS Service) standards supported by TEQC. Although the GNSS receiver Trimble R8 Model 2 and GNSS antenna TRIMBLE R8 GNSS/SPS88x are concerned here, due to correct data processing within TEQC, the receiver and antenna designations were harmonized with IGS standards: TRIMBLE R8 GNSS (receiver) and TRMR8_GNSS (antenna) [Figure 4.2]. Figure 4.2 Header of the edited RINEX file prepared for processing in TEQC with the observations gathered on the RGN station 4.1. TEQC Quality Check (QC) By using +qc options within TEQC, different quality indicators of input data are calculated and GNSS observation problems revealed. Along with observation data, the satellite ephemeris data (GPS and GLONASS) may be used optionally in order to improve QC results (qc-full) [URL 5]. GPS and GLONASS navigation files were used for data processing, so the results are given in qc-full format. TEQC's default mask elevation is 10 above the horizon. The horizon in TEQC is defined by the plane parallel to the ellipsoid through the receiver's antenna phase centre [ibid.]. The results of QC data processing are output in ASCII file with an S appended to the file extension (e.g. by processing of input file ZAGR053A.16o along with GPS and GLONASS navigation files, the output ZAGR053A.16S 259

6 SIG 2016 International Symposium on Engineering Geodesy, May 2016, Varaždin, Croatia ASCII file with qc-full results is generated). This file contains the processing report given in two parts: short report format and long report format. In the short format part, ASCII plot with graphical results of data processing for each individual satellite along with the summary report is presented, whereas in the second part, there are the details about, e.g. multipath per each satellite given, as well as the distribution of SN1 and SN2 values per elevation above horizon. The abbreviations SN1 and SN2 refer to the Signal to Noise Ratio (SNR) of phase observations on L1 and L2 carriers, respectively [Estey & Lou 2014]. The comparative analysis of QC processing results for the stations ZAGR and RGN is presented in sequel. In table 4.1, there is the information given about: receiver tracking capability (Rx tracking capability: 28 SV), number of possible epochs with observation (Poss. # of obs epoch: 8640), number of epochs with observation (Epoch w/ observations: 8640), possible observations above horizon (Possible obs >0.0 deg), possible observations above elevation mask 10 (Possible obs >10.0 deg), complete observations contained in observation (OBS) file above elevation mask (Complete obs >10.0 deg), deleted observations above elevation mask (Deleted obs >10.0 deg) and masked observations i.e. observations below the elevation mask 10 (Masked obs <10.0 deg). A complete observation within TEQC is related to the phase and code pseudorange data on both carriers (L1 and L2) for GPS and GLONASS data, and also "good" S/N data on both L1 and L2, and that is usually the case when the threshold is zero. Deleted observation are the observations deleted for any reason: below elevation mask, missing code or phase data, and/or poor S/N [URL 5]. Table 4.1 Number of possible, complete, deleted and masked observations gathered on stations ZAGR and RGN ZAGR RGN It is evident from the data in table 4.1 that GNSS receivers on the stations ZAGR (Trimble NetR5) and RGN (Trimble GNSS R8) were capable to track 28 satellites, there were 8640 possible observations using logging interval of 10 seconds (24 x 60 x 6 = 8640). Slightly greater number of possible observations above 10 at the ZAGR station (129686) compared to those on the RGN station (129645) are most likely due to the characteristics of the GNSS receiver and antenna set, bearing in mind that the horizon on the station RGN is free of 260

7 TS 2 Geodetic Networks, Data Quality Control, Testing and Calibration obstacles (see Chapter 5). The great difference in the number of complete observations between the stations ZAGR and RGN is due to the absence of code pseudoranges P2 (carrier L2) for 11 GPS satellites. Significantly greater number of masked observations on the RGN station is the direct effect of setting the elevation mask to 0 (elevation mask on the station ZAGR was set to 10 ). Table 4.2 Moving average (RMS) MP12 and MP21 as well as mean values of S1 and S2 ZAGR RGN TEQC computes the root mean square (RMS) of multipath combinations MP1 (MP12) and MP2 (MP21). The values MP1 and MP2 are TEQC's computed values of the RMS moving average values of the MP1 and MP2 in meters. There are 50 points contained in the moving average, and the observations above elevation mask 10 are taken into consideration. Moreover, the mean values of SNR1 and SNR2 are reported together with standard deviations and number of values the mean values are calculated from. The moving average RMS values of MP12 and MP21 along with S1 (SNR1) and S2 (SNR2) values for stations ZAGR and RGN are shown in table 4.2. Table 4.3 Number of observations on stations ZAGR and RGN along with percentage of observations above elevation mask without data and observations with low SNR values ZAGR RGN 261

8 SIG 2016 International Symposium on Engineering Geodesy, May 2016, Varaždin, Croatia It can be inferred from the data in table 4.2 that the moving average values MP12 and MP21 on the RGN station are in average 4 cm larger in comparison to those on the ZAGR station, the mean SNR1 values are approximately equal, the SNR2 values are higher on the RGN station. This may lead to the conclusion that the noise effect on the RGN station is smaller than on the ZAGR station, although it must be pointed out that GNSS observations on both stations were not gathered using equal receiver and antenna equipment. QC-full report includes a summary of MP12 and MP21 RMS values per each observed satellite as well as the mean value (RMS) for the station. The distribution of MP12 and MP21 values (RMS) per elevation intervals starting from down to 5-0 are graphically and numerically presented, showing that satellites at low elevations are significantly more burdened with multipath effect. On the RGN station, MP12 values are smaller for all elevation intervals greater than 50-55, MP21 values are smaller for all elevations. Graphical and numerical representation of S/N (SNR) distribution for carriers L1 and L2 per elevation intervals are given at the end of qc-full report. The total number of observations per elevation intervals along with mean S/N values and their standard deviation (1 sigma) are given too. Table 4.3 Comparative overview of S/N distribution on carrier L2 per elevation interval on stations ZAGR and RGN ZAGR RGN 262

9 TS 2 Geodetic Networks, Data Quality Control, Testing and Calibration The data in table 4.3 show that S/N mean values on L2 carrier are generally smaller on the ZAGR station. It is clearly visible that SNR values are increasing along with the elevation increase. Significantly more populated intervals 0-5 and <0 on the RGN station may be correlated with the elevation mask Baseline processing and network adjustment in TBC The ultimate goal of GNSS measurements is the determination of coordinates along with their accuracy estimation. The statically gathered observations on the RGN station together with the data downloaded for CROPOS CORS stations KARL (47.4 km), ZABO (25.1 km), SISA (47.6 km) and nearby ZAGR (0.115 km) were processed using Trimble Business Centre (TBC), ver For baseline processing, GPS and GLONASS broadcast as well as precise ephemerides were used. The precise GPS ephemerides (IGS final) were downloaded from URL 6, GLONASS precise ephemerides were downloaded from URL 7. Prior to baseline processing, project setting in TBC were defined as follows: coordinate system HTRS96/TM, GPS time, ephemerides broadcast/precise, usage of all satellites observations (GPS and GLONASS), elevation mask 10, antenna calibration NGS Absolute. After baseline processing using broadcast ephemeris, the network adjustment was performed with CORS KARL, ZABO, SISA and ZAGR coordinates fixed in reference frame ETRF2000 (R05) (official reference frame of CROPOS CORS network). The network adjustment after baseline processing using precise ephemeris given in IGS08 (IGS realization of ITRF2008 [Rebischung et al. 2012]) was carried out as follows: CORS coordinates were transformed from ETRF2000 (R05) (e = ) to ITRF2008 (e = ), network adjustment was performed in ITRF2008 with subsequent transformation back to ETRF2000 (R05) (e = ). The station velocities were estimated from model NNR Nuvel 1A using Plate Motion Calculator [URL 8]. The adjustment results with associated accuracy estimation are given with the confidence level 95%. Although the results of the baseline processing (using broadcast and precise ephemerides) and subsequent network adjustment have shown equal results, the solution derived with precise ephemerides was accepted as the final result. The final ellipsoid coordinates (GRS80) of the RGN station along with the coordinates of CROPOS CORS ZAGR are given in table 5.1. Table 5.1 Ellipsoid coordinates (GRS80) of stations RGN and ZAGR Station φ λ h [m] RGN 45 48' '' 15 57' '' ZAGR 45 48' '' 15 57' '' The estimation of coordinate accuracy is given as follows: ±0.003 m (E), ±0.002 m (N) and H ±0.003 m (h). These accuracy estimation values may be considered as additional confirmation of significant noise absence in GNSS observations on RGN station. The ellipsoid azimuth value from RGN to ZAGR station (α = ) has been calculated from coordinates in table 5.1. Using this 263

10 SIG 2016 International Symposium on Engineering Geodesy, May 2016, Varaždin, Croatia azimuth value and the total station Topcon GTS 212, horizontal directions and zenith distances to objects on the apparent horizon around RGN station were observed. It was found that on apparent horizon around RGN station there are no objects with elevation greater than 8.6 leading to the conclusion that such horizon with the application of elevation mask 10 can be considered totally free of obstacles. 6. Conclusion Although the suitability of site GNSS measurements may be assessed by means of terrain reconnaissance, a more reliable approach involves elevation angles determination of objects on the apparent horizon around the station. However, in order to get better insight about potential noise generated by the sources of electromagnetic radiation, it is necessary to gather a sufficient amount of statically observed data and subsequently process and analyze the results using appropriate software tool, e.g. TEQC. The quantities that have to be analysed are multipath and SNR on carriers L1 and L2 for GPS and GLONASS satellites as well. The increased multipath indicates the presence of reflective surfaces, whereas small SNR can indicate the potential presence of noise that comes from the sources of electromagnetic radiation. In spite of the presence of antenna and base station of mobile telecommunication providers in the proximity of the RGN station, the results of data processing in TEQC have not pointed out a significant source of noise. In comparison with QC results on CROPOS CORS ZAGR, the presence of slightly increased multipath on RGN station was revealed, but at the same time, there were higher SNR values for L2 carrier found. Since the observation on the stations RGN and ZAGR were carried out using different GNSS receiver and antenna combinations, more reliable analysis requires the observations to be made by using GNSS equipment of equal characteristics. The performance of GNSS measurements on station with horizon free of obstacles and the absence of noise generated by source of electromagnetic radiation are the main prerequisites for reliable determination of coordinates. References Bradke, B. (2009). Carrier-to-Noise Density and AI for INS/GPS Integration, InsideGNSS, Volume 4, Number 5, pp Estey, L.; Wier, S. (2014). Teqc Tutorial: Basics of Teqc Use and Teqc Products, UNAVCO, Boulder, Colorado U.S.A. Hofmann-Wellenhof, B.; Lichtenegger, H.; Wasle, E. (2008). GNSS Global Navigation Satellite System, Springer, Wien, New York. Joseph, A. (2010). Measuring GNSS Signal Strength, InsideGNSS, Volume 5, Number 8, pp

11 TS 2 Geodetic Networks, Data Quality Control, Testing and Calibration Rebischung, P.; Griffiths, J.; Ray, J.; Schmid, R.; Collilieux, H.; Garayt, B. (2012). IGS08: the IGS realization of ITRF2008, GPS Solutions, Volume 6, Issue 4, Springer Berlin Heidelberg. Seeber, G. (2003). Satellite Geodesy, 2nd completely revised and extended edition, Walter de Gruyter, Berlin, New York, pp Seepersad, G.; Bisnath, S. (2015). Reduction of PPP convergence period through pseudorange multipath and noise mitigation, GPS Solutions, Volume 19, Issue 3, Springer-Verlag, Berlin, Heidelberg, pp Trimble Navigation Limited (2013). Trimble Survey Controller Help, Version 12.50, Revision A, Engineering Construction Group, Dayton, Ohio, U.S.A. Zrinjski, M. (2010). Defining the Calibration Baseline Scale of the Faculty of Geodesy by Applying Precise Electro-optical Distance Meter and GPS (in Croatian), doctoral dissertation, Faculty of Geodesy, University of Zagreb, Zagreb. URL 1: MP1 and MP2 derivation, ( ). URL 2: CROPOS GNSS reference stations web server, ( ). URL 3: Receiver Independent Exchange Format Version 2.11, ( ). URL 4: Trimble NetR5 GNSS User Guide, /NetR5_UserGde_3.10A_ENG_Web.pdf, ( ). URL 5: TEQC Tutorial, ( ). URL 6: IGS Product Availability, ( ). URL 7: IGS Product, ftp://cddis.gsfc.nasa.gov/pub/glonass/products/1885/, ( ). URL 8: Plate Motion Calculator, ( ) 265

12 SIG 2016 International Symposium on Engineering Geodesy, May 2016, Varaždin, Croatia Ispitivanje pogodnosti lokacije za GNSS mjerenja Sažetak. U svrhu provedbe kvalitetnih GNSS opažanja koja će rezultirati pouzdanim određivanjem koordinata točaka, posebice onih koje se često ili trajno koriste, potrebno je da lokacija na kojoj će se izvoditi mjerenja zadovolji određene kriterije što ponajprije uključuje horizont bez zapreka, ali i izostanak šuma generiranih izvorima elektromagnetskog zračenja. Ispunjavanje navedenih kriterija ispitano je na osnovi podataka prikupljenih tijekom 24-satne sesije GNSS opažanja na točki na vrhu tornja Rudarsko-geološko-naftnog fakulteta (RGN) ispod kojeg se nalaze antene i bazna stanica operatera mobilne telekomunikacijske mreže. Prikupljeni podaci opažanja obrađeni su pomoću programa TEQC, a dobivene su informacije o SNR-u (engl. Signal-To-Noise Ratio) kao i o višestrukoj refleksiji signala (engl. multipath) na L1 i L2 frekvencijama kako GPS tako i GLONASS satelita. Slično tome, obrađena su i opažanja s obližnje CROPOS CORS ZAGR te su prikazani dobiveni rezultati. Na osnovi GNSS opažanja prikupljenih na točki RGN zajedno s opažanjima na CROPOS referentnim stanicama KARL, SISA, ZABO i ZAGR izračunani su vektori, provedeno je izjednačenje mreže te je dana ocjena točnosti dobivenih koordinata. Dodatno, pomoću totalne mjerne stanice opažan je prividni horizont oko točke RGN iz čega se zaključilo da je horizont potpuno čist iznad elevacijske maske 10. Ključne riječi: GNSS opažanja, horizont, omjer signal-šum (SNR), TEQC, višestruka refleksija signala (multipath). *scientific paper 266