J. Geod. Sc. 2014; 4:65 73 Journal of Geodetc Scence Research Artcle Open Access J. Tegedor, O. Øvstedal, and E. Vgen Precse orbt determnaton and pont postonng usng GPS, Glonass, Galleo and BeDou Abstract: State of the art Precse Pont Postonng PPP s currently based on dual-frequency processng of GPS and Glonass navgaton systems. The Internatonal GNSS Servce IGS s routnely provdng the most accurate orbt and clock products for these constellatons, allowng pont postonng at centmeter-level accuracy. At the same tme, the GNSS landscape s evolvng rapdly, wth the deployment of new constellatons, such as Galleo and BeDou. The BeDou constellaton currently conssts of 14 operatonal satelltes, and the 4 Galleo In-Orbt Valdaton IOV satelltes are transmttng ntal Galleo sgnals. Ths paper focuses on the ntegraton of Galleo and BeDou n PPP, together wth GPS and Glonass. Satellte orbts and clocks for all constellatons are generated usng a network adjustment wth observaton data collected by the IGS Mult-GNSS Experment MGEX, as well as from Fugro propretary reference staton network. The orbt processng strategy s descrbed, and orbt accuracy for Galleo and BeDou s assessed va orbt overlaps, for dfferent arc lengths. Knematc post-processed mult-gnss postonng results are presented. The benefts of multconstellaton PPP are dscussed n terms of enhanced avalablty and postonng accuracy. Keywords: BeDou, Galleo, Intersystem-bases, Multconstellaton Precse Pont Postonng DOI 10.2478/jogs-2014-0008 Receved January 20, 2014 ; accepted Aprl 3, 2014. 1 Introducton The Precse Pont Postonng PPP technque Zumberge et al. 1997 has become ncreasngly sgnfcant n hgh-precson postonng applcatons durng recent J. Tegedor: Norwegan Unversty of Lfe Scences NMBU, Ås, Norway, E-mal: javer.tegedor@nmbu.no O. Øvstedal: Norwegan Unversty of Lfe Scences NMBU, Ås, Norway E. Vgen: Fugro Satellte Postonng AS, Oslo, Norway years Kanzak et al. 2011; Geng et al. 2010, as t allows the estmaton of accurate recever coordnates, wthout the need of a nearby reference staton. PPP has other nterestng applcatons, such as tme-transfer Defragne et al. 2008, onospherc Leandro et al. 2007 and tropospherc characterzaton Kjørsvk et al. 2006, or bases calbraton Leandro et al. 2010. The Internatonal GNSS Servce IGS s routnely generatng the most accurate orbt and clock estmates, for GPS and Glonass satelltes, by means of a dense global network and several contrbutng analyss centers Dow et al. 2009. Makng use of these products and precse observaton modelng Kouba and Héroux 2010, statc absolute postonng can be acheved at centmeter level accuracy n post-processng. Sub-decmeter level accuracy can be acheved n knematc applcatons Hesselbarth 2011. Real-tme users can also access orbt and clock correctons va RTCM streams, enablng decmeter-level accuracy n real-tme Cassy et al. 2012. In addton, there are several commercal PPP servces makng use of GPS and Glonass, such as Fugro s G2 Melgard et al. 2009 or Trmble s RTX Leandro et al. 2011, whch also supports the Japanese QZSS Quas-Zenth Satellte System. The current development of BeDou and Galleo constellatons offers new prospects for precse navgaton, when combned wth tradtonal GPS and Glonass PPP, thanks to the ncreased number of satelltes avalable. At the tme of wrtng, the BeDou constellaton conssts of 5 Geostatonary Orbt GEO, 5 Inclned Geosynchronous Orbt IGSO and 4 Medum-Earth Orbt MEO satelltes, provdng regonal coverage around Chna for contnuous postonng. The constellaton deployment s expected to resume n 2014, wth the further development of the MEO constellaton, n order to acheve global coverage before the end of ths decade. The Galleo constellaton s currently composed of 4 ntal In-Orbt-Valdaton IOV satelltes. The Full Operatonal Capablty FOC phase s expected to start also n 2014 wth the launch of the frst operatonal satelltes. Ths artcle focuses on the contrbuton of Galleo and BeDou to PPP. A prerequste s the generaton of precse 2014 J.Tegedor et al., lcensee De Gruyter Open. Ths work s lcensed under the Creatve Commons Attrbuton-NonCommercal-NoDervs 3.0 L- Unauthentcated cense.
66 J. Tegedor, O. Øvstedal, and E. Vgen satellte orbts and clocks for the new constellatons. Ths s done usng a network least-squares adjustment, makng use of observaton data from the IGS Mult-GNSS Experment MGEX, as well as from Fugro s propretary network, whch has been made avalable for ths study. The network and the observables avalable are descrbed n secton 2. Ionosphere-free observaton equatons for mult- GNSS PPP are presented n secton 3, where constellatondependent ntersystem bases are ntroduced, for Galleo and BeDou. The processng strategy for orbt and clock estmaton s presented n secton 4, together wth an assessment of the orbt accuracy. The estmated ntersystem bases are presented n secton 5. Knematc mult-gnss postonng results are presented n secton 6, where the benefts of addng Galleo and BeDou to PPP are dscussed. Conclusons are summarzed n secton 7. 2 Trackng data For GPS and Glonass, the Internatonal GNSS Servce IGS has been provdng observaton data for scentfc purposes durng the last twenty years. Wth the development of new navgaton systems, IGS started n 2011 the Mult- GNSS Experment MGEX Rzos et al. 2013, amng at upgradng the current staton network to support new constellatons. Most of the statons n the MGEX campagn are Galleo-capable, and a subset of them are also trackng BeDou. The network also observes the Japanese QZSS system, but ths system has not been ncluded n ths study, as the contrbuton to PPP s stll relatvely small, wth only one satellte QZSS-1 avalable at the moment. At the same tme, Fugro s operatng a worldwde reference staton network for supportng ts commercal postonng servces, manly for martme applcatons. A subset of the statons Fg. 1. Geographcal dstrbuton of MGEX and Fugro statons, ndcatng trackng capablty for Galleo and Bedou August 2013. n the network has been upgraded to Galleo and BeDou capablty. A map of the statons avalable n both networks s depcted n Fg. 1. It can be observed that, although there s a concentraton of statons n Europe, the network stll provdes a farly good global coverage. Partcularly relevant are the statons located n Asa-Pacfc area, for the trackng of BeDou IGSO and GEO satelltes. Regardng the frequency plan for the new constellatons, BeDou s broadcastng sgnals n three carrer frequences: 1589.74 MHz B1, 1207.14 MHz B2 and 1268.52 MHz B3 Greler 2007, whereas Galleo s transmttng open sgnals n E1 1575.42 MHz, E5a 1176.45 MHz, E5b 1207.14 MHz and E5a+b 1195.795 MHz European Commson 2010, n addton to the future commercal servce n the E6 carrer 1278.75 MHz. It s to be noted that BeDou and Galleo are sharng the B2/E5b carrer, whle Galleo and GPS are sharng both L1/E1 and L5/E5a carrers. Ths opens the door to nteroperablty applcatons, whch are out of the scope of ths study. The reader s referred to Melgard et al. 2013 for a study on nteroperablty of GPS and Galleo usng E1/E5a n PPP. All the Fugro statons, equpped wth Trmble NETR9 recevers, are provdng Galleo and BeDou data n all frequences. However, t has been found that, for MGEX statons, the sgnals avalablty depends on the recever model, the frmware nstalled n each recever and/or the way of generatng Rnex3 fles from raw data. Table 1 summarzes the number of statons per recever type n the network, and the avalablty of Galleo and BeDou trackng for each recever type. For nstance, the majorty of Javad G3T Delta recevers are trackng Galleo only on E1 and E5a frequences. Actually only one Javad recever WTZZ, Wetzell, Germany, whch s equpped wth the most modern recever board and frmware, s also trackng E5b and E5a+b sgnals, as well as BeDou B1 and B2 sgnals. Septentro recevers are generally not able to track the BeDou B3 sgnals, due to a lmtaton n the current recever frmware. Some other statons equpped wth Trmble NETR9 recevers are not provdng any Galleo or Be- Dou measurements, probably due to the way these recevers are confgured by the staton operators. In order to maxmze data avalablty wth the exstng observatons, Galleo E1 and E5a, as well as BeDou B1 and B2 sgnals have been selected for the subsequent analyss. The observables are processed usng the onosphere-free lnear combnaton, whose observaton equatons are detaled n the next secton. Unauthentcated
Precse orbt determnaton and pont postonng usng GPS, Glonass, Galleo and BeDou 67 Table 1. Recever type dstrbuton n the MGEX network as per August 15 th, 2013. Number of statons trackng each Galleo and BeDou frequency are shown. Recever Brand Model Number of statons Galleo BeDou E1 E5a E5b E5a+b B1 B2 B3 Javad Delta G2T 3 3 3 0 0 0 0 0 Delta G3T 23 23 23 1 1 1 1 0 GR10 4 4 4 4 4 0 0 0 Leca GR25 4 4 4 1 1 0 0 0 GRX1200 5 5 5 5 5 0 0 0 Novatel OEM6 1 1 1 0 0 0 0 0 AsteRx3 1 1 1 1 1 1 1 0 Septentro PolaRx4 6 6 6 6 6 6 6 0 PolaRx4TR 2 2 2 2 2 1 1 0 PolaRxS 1 1 1 1 1 1 1 0 Trmble NETR9 26 22 22 22 22 17 17 17 TOTAL 76 72 72 43 43 27 27 17 3 Observatons equatons For ths study, the GPS observaton equatons proposed n Collns et al. 2008 have been extended for accommodatng mult-gnss observatons, addng nter-systembas parameters between dfferent constellatons. The resultng onosphere-free observatons equatons, for each GNSS, for pseudorange P and carrer-phase L, between staton and satellte j, are: P j,gps L j,gps P j,glo L j,glo P j,gal L j,gal P j,bei L j,bei where: = ρ j + νj Γ + c δt δt j + ε P 1 = ρ j + νj Γ + c δt δt j + a j + ε L 2 = ρ j + νj Γ + c = ρ j + νj Γ + c = ρ j + νj Γ + c = ρ j + νj Γ + c = ρ j + νj Γ + c = ρ j + νj Γ + c δt δt j + ISB j,glo δt δt j + ISB j,glo δt δt j + ISB GAL δt δt j + ISB GAL δt δt j + ISB BEI δt δt j + ISB BEI + ε j P 3 + a j + εj L 4 + ε j P 5 + a j + ε j L 6 + ε j P 7 + a j + εj L 8 ρ j s the geometrc dstance between staton and satellte, assumng relevant correctons, such as antenna phase center correctons or phase wnd-up, have been already accounted for. ν j Γ s the wet tropospherc delay between staton and satellte, where Γ s the tropospherc zenth delay and ν j s the assocated elevaton-dependng mappng functon. The dry component of the tropospherc delay s removed from the observatons usng an a-pror model. δt and δt j are the epoch-wse recever and satellte clock offsets, respectvely. Followng ths notaton, t has been assumed that there s a sngle recever clock common to all observables from dfferent GNSS. GNSSdfferences are accounted for n the ntersystem-bas terms. c s the speed of lght. ISB j,glo s the GPS-Glonass ntersystem-bas term. It s to be noted that ths bas depends on each staton and satellte j, due to the Frequency Dvson Multple Access FDMA scheme mplemented by Glonass, whch nduces recever- and satellte-dependent nterchannel bases. As shown n Reussner and Wannnger 2011, each frequency satellte encounters a slghtly dfference delay n the recever. ISB GAL and ISB BEI are the GPS-Galleo and GPS- BeDou ntersystem-bases, respectvely. Contrary to Glonass, t s to be noted that these are satelltendependent, as Galleo and BeDou have adopted the Code Dvson Multple Access CDMA scheme, meanng that all satelltes from the same constellaton use the same carrer-frequency. Sgnfcant bases appear though dependng on the recever model, Unauthentcated
68 J. Tegedor, O. Øvstedal, and E. Vgen but mght depend also on the dgtal sgnal processng frmware happenng nsde the recever. It needs to be mentoned that the new generaton of Glonass satelltes Glonass-K s expected to mplement CDMA as well. a j s the ambguty term between staton and satellte, assocated to the carrer-phase measurements. For the onosphere-free lnear combnaton, ths term s n general not nteger, due to non-nteger nature of the combnaton coeffcents, and the presence of satellte and recever hardware delays Laurchesse et al. 2009. ε j P and ε j L are unmodelled effects, such as thermal nose and multpath, for pseudorange and carrerphase, respectvely. It s to be noted that, for GNSS measurements, ε j L ε j P. 4 Orbt and clock estmaton 4.1 Processng strategy For estmaton of orbt and clocks, NAPEOS Sprnger and Dow 2009 software package has been used. The software has been extended to process BeDou, on top of the exstng capabltes for GPS, Glonass and Galleo. The processng strategy, depcted n Fg. 2, wll be descrbed next. In order to obtan an a-pror orbt, broadcast ephemers can be used for GPS and Glonass. For Galleo, test ephemers started n March 2013, but satelltes are stll unhealthy meanng that ths data mght not always be relable. For BeDou, MGEX statons are at the moment not provdng any ephemers. For these reasons, Two Lne Elements TLEs are used, for both Galleo and BeDou, n order to obtan an a-pror orbt ntalzaton. The accuracy of ths a-pror orbt s at sub-klometer level. TLEs can be downloaded from www.space-track.org, whch also ncludes the format descrpton. In a frst processng step, a least-squares estmaton usng only pseudorange observatons s performed, n order to mprove the TLE-derved orbts and to obtan a-pror satellte clocks for Galleo and BeDou. After ths step, the orbt accuracy s around meterlevel, smlar to what s obtaned for GPS and Glonass va broadcast ephemerdes. In a second step, both pseudorange and carrer-phase observatons are used, n order to beneft from the precson of the carrer-phase measurements. Estmated parameters are the satellte state vectors, solar radaton pressure parameters, wet tropospherc delays, satellte and staton Fg. 2. Processng strategy for generaton of orbt and clocks, ncludng Galleo and BeDou. clocks, ntersystem-bas terms and carrer-phase ambgutes. Fnally, n order to obtan sutable clocks for PPP at 30 seconds samplng, a fnal clock densfcaton process s performed. In ths fnal step, only staton and satellte clocks are estmated. All other parameters are kept fxed to the prevous estmates. 4.2 Modelng for Galleo and BeDou A summary of the models beng used for all constellatons s presented n Table 2. Beng relatvely new constellatons, BeDou and Galleo have a number of modelng lmtatons compared to more mature systems, such as GPS and Glonass. The mpact of these lmtatons wll be addressed n the current secton. GPS and Glonass precse antenna phase center correctons Schmd et al. 2007 have been made avalable as part of the IGS actvtes va the Antenna Exchange AN- TEX format, both for transmttng and recevng antennas. These precse correctons are not avalable yet for ether Galleo or BeDou. For satellte antennas, the MGEX project has released approxmate values for the dstance between the satellte center of mass and the antenna phase center. These are [0.2, 0, 0.6] m for Galleo and [0.6, 0, 1.1] m for BeDou, XYZ n the satellte body-fxed reference frame. It Unauthentcated
Precse orbt determnaton and pont postonng usng GPS, Glonass, Galleo and BeDou 69 Table 2. Summary of models used for mult-gnss processng. GPS Glonass Galleo BeDou Observaton samplng 30 seconds Elevaton cut-off 10 degrees Sgnal selecton L1/L2 E1/E5a B1/B2 Antenna phase-center correctons IGS Antex fle A-pror values Tropospherc modelng GPT/GMF Boehm et al. 2007 Ionospherc modelng Frst order removed by lnear combnaton Solar Radaton Pressure CODE Emprcal Model wth 5 parameters s expected that these values have an uncertanty around decmeter level. Nadr- or azmuth-dependent correctons are not avalable for these constellatons so far. On the recevng antennas, the phase center offset and azmuth- and elevaton-dependent varatons for Galleo and BeDou frequences are expected to be slghtly dfferent up to few centmeters to the ones used for GPS, due to the dfferent frequences used by Galleo and BeDou. At the tme of wrtng, there are no publcly avalable calbratons for the antennas used n MGEX statons. For ths study, GPS calbratons have been used for Galleo and Be- Dou, whch ntroduces an addtonal uncertanty below decmeter level. Satellte atttude modelng s not fully known for the new constellatons. Under nomnal atttude, yaw-steerng mode has been assumed for Galleo and BeDou, n the same way as for GPS Kouba 2008. The behavor of the new satelltes under eclpse seasons remans a topc for further research. A msmodellng of the satellte atttude n GNSS mpacts the wnd-up correcton n carrer-phase measurements Wu et al. 1993, due to the relatve orentaton between transmttng and recevng antennas. In order to lmt the mpact of ths uncertanty n PPP, the very same models have been appled n the orbt adjustment and n the precse pont postonng estmaton. 4.3 Orbt qualty For the orbt estmaton, two dfferent sets of daly solutons have been generated, n order to assess the mpact of the orbt arc length on the orbt accuracy. The frst soluton s based on 24 hours arcs, whle the second s based of 72 hours, where the central 24 hours are extracted as daly solutons. In order to estmate the orbt qualty for GPS and Glonass, a comparson wth IGS Fnal products has been performed. For the 3-day arc orbts, the monthly RMS s 1.7 cm for GPS, 3.9 cm for Glonass. Regardng BeDou and Galleo satelltes, the orbt precson can be assessed measurng orbt dfferences between consecutve solutons at day boundares. The monthly RMS values obtaned for these day boundary dfferences, for both 1-day and 3-days arc solutons, for each Galleo and BeDou satellte, are shown n Table 3. Sample GPS and Glonass satelltes have been ncluded for reference. It can be observed that 3-day orbt soluton mproves sgnfcantly the orbt precson wth respect to the 1-day soluton, thanks to the better observablty of the orbt dynamcs over longer data arcs. Addtonally, GEO orbt precson s typcally lower than MEO and IGSO orbts, manly on the along-track component. Ths could be explaned by the fact that there s no geometry varaton between the GEO satelltes and the reference staton network, whch weakens the orbt estmaton. Nevertheless, sub-decmeter level accuracy could stll be acheved on the radal and cross-track components. It s also nterestng to note that the 3-day orbt solutons for IGSO satelltes C07 and C10 are sgnfcantly worse than other IGSO satelltes. The reason beng that these satelltes were under Earth eclpse perods durng the frst two weeks of August 2013. As mentoned earler, precse atttude modelng for these satelltes needs to be studed n order to obtan accurate orbts also durng eclpse seasons. For Precse Pont Postonng, the 3-day orbt soluton wll be used, n order to acheve the hghest possble accuracy. Addtonally, the GEO satelltes have been deweghted wth a factor of 3 n PPP wth respect to IGSO and MEO, n order to account for orbt uncertanty n these satelltes. 5 Intersystem bases When processng mult-gnss observatons, ntersystembases need to be taken nto account. For Glonass, these Unauthentcated
70 J. Tegedor, O. Øvstedal, and E. Vgen Table 3. Orbt day-boundary dfferences RMS, durng August 2013, for 24 and 72 hours orbt arcs. All unts are centmeters. Satellte Type GPS GLONASS Galleo BeDou GEO BeDou IGSO BeDou MEO PRN 1-day orbt arcs 3-day orbt arcs Radal Along-track Cross-track Radal Along-track Cross-track G01 4.5 4.6 2.7 0.6 0.5 0.5 G25 1.7 5.3 2.3 0.4 0.5 0.5 R02 3.1 7.7 4.4 0.3 1.1 1.3 R03 3.1 8.1 4.7 0.4 1.3 1.2 E11 7.4 18.3 9.8 2.7 11.6 3.4 E12 6.0 15.0 9.5 3.5 8.7 2.3 E19 4.8 22.5 12.3 1.3 3.7 3.5 E20 4.2 19.2 9.3 1.2 3.7 3.5 C01 16.2 87.3 15.2 2.0 24.2 3.9 C02 81.9 185.9 9.0 6.3 30.7 11.9 C03 45.8 121.0 19.3 10.4 38.7 5.1 C04 34.4 76.7 10.6 3.3 25.9 5.2 C05 50.1 113.7 15.3 5.3 30.7 9.1 C06 58.9 22.4 16.9 3.0 4.3 2.7 C07 24.7 22.1 24.3 15.4 60.5 11.8 C08 22.8 12.8 16.6 1.8 4.7 2.5 C09 17.0 10.9 10.2 1.0 2.2 1.8 C10 24.6 15.3 17.6 13.1 35.1 9.6 C11 4.5 44.8 11.5 1.2 6.6 2.6 C12 4.2 54.3 10.6 1.4 7.3 2.8 C13 5.0 48.5 16.2 0.9 7.9 4.8 C14 5.2 50.2 17.7 0.9 8.2 5.2 bases have been extensvely analyzed n the lterature Chuang et al. 2013; Wannnger 2012, and ths study wll manly focus on Galleo and BeDou. As descrbed n secton 3, a sngle parameter per staton and system ether Galleo or BeDou s enough to account for ntersystem bases, as all satelltes are usng the same carrer frequency. Addtonally, n order to defne the clock datum, a zero-mean condton has been appled to all ntersystembases n the orbt and clock estmaton process, for each constellaton. Ths approach allows assessng relatve dfferences n ntersystem-bases between dfferent recevers n the network. The daly ntersystem bases for each constellaton are depcted n Fg. 3. Generally, a strong recever-type dependency can be observed, wth all statons wth the same recever model showng smlar bases. An excepton has been found n recever WTZZ Javad Delta G3T, whch shows sgnfcant dfferences wth respect to other Javad recevers. One possble explanaton s that ths recever has a dfferent archtecture that allows t to track also Be- Dou. It s also to be noted that the BeDou ntersystembas for WTZZ s dfferent by more than 100 ns compared to Trmble or Septentro recevers, meanng the effect s sgnfcant enough and cannot be gnored for precse applcatons. Addtonally, t s notceable that there are stll small remanng dfferences wth statons equpped wth the same recever type. Ths mght be due to antennaor cable-nduced delays, or thermal effects between hardware nstallatons at dfferent locatons. Fg. 3. Intersystem-bas estmates for several recevers n the MGEX network, for Galleo top and BeDou bottom. Unauthentcated
Precse orbt determnaton and pont postonng usng GPS, Glonass, Galleo and BeDou 71 Fg. 4. MGEX statons selected for mult-gnss PPP. 6 Precse Pont Postonng assessment For mult-constellaton Precse Pont Postonng, a new knematc PPP algorthm has been mplemented n NAPEOS, based on sequental least squares, followng the gudelnes gven n Kouba 2009, and the observaton equatons descrbed n secton 3. In order to assess the effect of mult-constellaton precse pont postonng, some reference statons from the MGEX network have been selected. These are KOUR Kourou, French Guyana, BRST Brest, France, NNOR New Norca, Australa and JFNG Jufeng, Chna. The staton locatons are ndcated n Fg. 4. All statons are trackng GPS, Glonass, Galleo and Be- Dou MEO. In addton, JFNG and NNOR are also trackng BeDou IGSO and GEO satelltes, thanks to ther geographcal locaton. Fgure 5 shows knematc mult-gnss PPP results for staton NNOR on August 26 th, 2013. The 95% poston error quantles are 1.74, 1.16 and 3.95 cm, n the East, North and Up components, respectvely, after removng the frst two hours of convergence perod. It s nterestng to notce the hgh number of satelltes avalable for PPP when usng all 4 constellatons, resultng n a very stable geometry Dluton of Precson-DOP. In order to assess the benefts of mult-gnss PPP, daly knematc PPP results have been obtaned for the month of August 2013, for all four statons, n dfferent confguratons: GPS only, GPS+Glonass, GPS+Glonass+Galleo and GPS+Glonass+Galleo+BeDou. The monthly average of the daly 95% poston error percentle s summarzed n Fg. 6. As shown n prevous studes Hesselbarth 2011, the contrbuton of Glonass on top of GPS s qute sgnfcant n terms of knematc postonng, thanks to the ncreased number of satelltes avalable and mproved geometry. For example, the NNOR vertcal error s reduced Fg. 5. PPP Knematc results for staton NNOR on August 26 th, 2013. by 36.3% when addng Glonass on top of GPS. Galleo further mproves the vertcal error by 3.5%, and the addtonal mprovement wth BeDou s 6.7%. The contrbuton of Galleo on top of GPS and Glonass s relatvely small, due to the small number of satelltes avalable, whch s vsble from a staton for lmted hours per day. The contrbuton of BeDou s slghtly more mportant, especally n JFNG and NNOR, where IGSO and GEO also contrbute to the postonng soluton n those locatons. Mult-constellaton Fg. 6. Mult-constellaton knematc postonng statstcs for several statons n MGEX. PPP s partcularly sutable n stuatons wth reduced sky Unauthentcated
72 J. Tegedor, O. Øvstedal, and E. Vgen vsblty, where the ncreased number of satelltes allows to obtan a sgnfcant hgher avalablty and accuracy compared to standalone GPS. In order to smulate ths scenaro, the PPP engne has been run wth several elevaton cut-off angles from 0 full sky vsblty to 35 reduced sky vsblty degrees. The results for staton NNOR are depcted n Fg. 7, n terms of postonng accuracy and average dluton of precson. It can be observed how the accuracy of the GPS-only soluton degrades rapdly wth partal sky vsblty. The mult-gnss soluton behaves sgnfcantly better n ths condton, n partcular the one wth all four constellatons, where sub-decmeter level accuracy can stll be obtaned even n the 35 degrees cut-off scenaro, manly thanks to the ncreased number of satelltes vsble for BeDou, on top of GPS, Glonass and Galleo. In ths case, Galleo mproves the vertcal accuracy by 12.6%, and BeDou brngs an addtonal 33.1% mprovement. Fg- be acheved when Galleo and BeDou MEO constellatons wll be fully deployed. Regardng statc PPP, t has been found that the addton of Galleo and BeDou does not sgnfcantly mprove the daly coordnate repeatablty. The reason s that the qualty of 24 hours GPS-only PPP s already at centmeter level, and the addton of new constellatons does not mprove sgnfcantly the accuracy. Ths was shown already for the case of Glonass n Hesselbarth 2011. Fg. 8. PPP avalablty for statons NNOR and JFNG on August 26 th, 2013. 7 Summary and Conclusons Fg. 7. Postonng statcs and Dluton of Precson DOP, for dfferent statons and cut-off angles, for staton NNOR on August 26 th, 2013. ure 8 represents the poston avalablty for statons JFNG and NNOR wth dfferent elevaton cut-off angles. Poston avalablty s here defned by the percentage of tme when 5 or more satelltes are vsble wth a Geometrc DOP GDOP lower than 10. The soluton wth all four systems s sgnfcantly better. It needs to be mentoned that these results correspond to statons where BeDou IGSO and GEO are avalable. In other regons of the world, the mprovement s currently not so sgnfcant. A worldwde contrbuton wll The GNSS landscape s evolvng rapdly, wth the addton of emergng satellte systems on top of GPS and Glonass. In ths study, precse orbt estmaton results have been presented for Galleo and BeDou. The acheved orbt precson s generally at sub-decmeter level for Galleo and Be- Dou MEO and IGSO satelltes. The orbt estmaton for GEO satelltes s challengng due to the lack of geometry varaton wth respect of the reference staton network, and precson estmates are at few decmeter level. Satellte modelng remans an area for further research, n terms of antenna phase center correctons and precse atttude modellng. Sgnfcant ntersystem-bases dfferences have been detected between dfferent recever brands, whch cannot be neglected for precse applcatons. Extended observaton equatons have been presented to accommodate these bases, both n network adjustment and PPP solutons. Mult-GNSS PPP knematc results show enhanced accuracy when usng all four satellte systems together. However, the accuracy mprovement s relatvely small compared to the GPS+Glonass case under good sky vsbl- Unauthentcated
Precse orbt determnaton and pont postonng usng GPS, Glonass, Galleo and BeDou 73 ty. The mprovement becomes more sgnfcant under reduced sky vsblty condtons, where the ncreased number of satelltes allows obtanng sgnfcantly hgher accuracy and avalablty for the poston soluton. Ths s partcularly vsble n the Asa-Pacfc area where BeDou IGSO and GEO satelltes are avalable for postonng. It can be expected that ths level of performance wll be extended worldwde wth the further deployment of Galleo and Be- Dou durng ths decade. In ths context, the data provded by the IGS MGEX campagn s hghly valuable for the scentfc communty to get a better understandng of the new GNSS systems and sgnals. Ths study would not have been possble wthout such data. The authors are also grateful to Fugro for delverng the reference staton data for scentfc purposes. References Zumberge J.F., Hefln M.B, Jefferson D.C. and Watkns M. M., 1997, Precse pont postonng for the effcent and robust analyss of GPS data from large networks, J Geophys. Res., 102, 1, 5005 5017. Kanzak M., Matsushta Y. and Kakmoto H., 2011, GNSS Postonng of Ocean Buoys n Japan for Dsaster Preventon, Proc. Int. Tech. Meetng of the Satelte Dvson the Insttute of Navgaton, 717 723. Geng J., Teferle F., Meng X. and Dodson A., 2010, Knematc precse pont postonng at remote marne platforms, GPS Solut, 14, 343 350. Defragne P., Guyennon N. and Bruynnx C., 2008, GPS Tme and Frequency Transfer: PPP and Phase-Only Analyss, Int. J. of Navgaton and Observaton, 1 7. Leandro R., Santos M. and Langley R., 2007, PPP-based onospherc actvty montorng, Proc. Int. Tech. Meetng of the Satellte Dvson of The Insttute of Navgaton, 2849 2853. Kjørsvk N., Gjevestad J. and Øvstedal O., 2006, Handlng of the tropospherc delay n knematc precse pont postonng, Proc. Int. Tech. Meetng of the Satelte Dvson of the Insttute of Navgaton, 2279 2285. Leandro R., Santos M. and Langley R., 2010, Analyzng GNSS data n precse pont postonng software, GPS Solut, 15, 1 13. Dow J., Nelan R. and Rzos C., 2009, The Internatonal GNSS Servce n a changng landscape of global navgaton satellte systems, J Geod, 83, 3, 191 198. Kouba J. and Héroux P., 2001, GPS Precse Pont Postonng usng IGS orbt products, GPS Solut, 5, 2, 12 28. Hesselbarth A., 2011, Statsche und knematsche GNSS- Auswertung mttels Precse Pont Postonng PPP. PhD Thess, Technschen Unverstät Dresden. Cassy M., Agrots L., Weber G., Hernandez-Pajares M. and Hugentobler U., 2012, The Internatonal GNSS Real-tme Servce, GPS World, June 2012. Melgard T., Vgen E., de Jong K. and Oerpen O., 2009, G2 - The Frst Real-Tme GPS and GLONASS Precse Orbt and Clock Servce, Proc. of the Int. Tech. Meetng of the Satellte Dvson of the Insttute of Navgaton, 1885 1891. Leandro R., Landau H. and Ntschke M., 2011, RTX Postonng: The Next Generaton of cm- accurate Real-Tme GNSS Postonng, Proc. of the Int. Tech. Meetng of The Satellte Dvson of the Insttute of Navgaton, 1460 1475, 2011. Rzos C., Montenbruck O., Weber R., Weber G., Nelan R. and Hugentobler U., 2013, The IGS MGEX Experment as a Mlestone for a Comprehensve Mult-GNSS Servce, Proc. of the ION 2013 Pacfc PNT Meetng, Honolulu, Hawa, 289 295. Greler T., 2007, Intal Observatons and analyss of Compass MEO Satellte Sgnals, Insde GNSS, May/June, 39 42. European Commson, 2010, European GNSS Galleo Open Servce Sgnal n Space Interface Control Document, Tech. report. Melgard T., Tegedor J., de Jong K., Lapucha D. and Lachapelle G., 2013, Interchangeable Integraton of GPS and Galleo by Usng a Common System Clock n PPP, Proc. of the Int. Tech. Meetng of the Satellte Dvson of the Insttute of Navgaton, 1198-1206. Collns P., Bsnath S., Lahaye F., and Héroux P., 2008, Undfferenced GPS Ambguty Resoluton usng the Decoupled Clock Model and Ambguty Datum Fxng, Proc. of the Int. Tech. Meetng of the Satellte Dvson of the Insttute of Navgaton, 1315 1322. Reussner N. and Wannnger L., 2011, GLONASS Inter-frequency Bases and Ther Effects on RTK and PPP Carrer-phase Ambguty Resoluton, Proc. of the Int. Tech. Meetng of the Satellte Dvson of the Insttute of Navgaton, 712 716. Laurchesse D., Mercer F., Berthas J. and Broca P., 2009, Integer ambguty resoluton on undfferenced GPS phase measurements and ts applcaton to PPP and satellte precse orbt determnaton, J Inst Navg, 56, 2, 135 149. Sprnger T. and Dow J., 2009, NAPEOS Mathematcal Models and Algorthms 1.0, European Space Agency, DOPS-SYS-TN-0100- OPS-GN. Boehm J., Henkelmann R. and Schuh H., 2007, Short Note: A global model of pressure and temperature for geodetc applcatons J Geod, 81, 679 683. Schmd R., Stegenberger P., Gendt G., Ge M. and Rothacher M., 2007, Generaton of a consstent absolute phase-center correcton model for GPS recever and satellte antennas, J Geod, 81, 781 798. Kouba J., 2008, A smplfed yaw-atttude model for eclpsng GPS satelltes, GPS Solut, 13, 1 12. Wu J., Wu S., Hajj G., Bertguer W., and Lchten S., 1993, Effects of Antenna Orentaton on GPS Carrer Phase Measurements, Man. Geod., 18, 91 98. Chuang S., Wentng Y., Wewe S., Ydong L. and Ru, Z., 2013, GLONASS pseudorange nterchannel bases and ther effects on combned GPS/GLONASS precse pont postonng, GPS Solut, 17, 439 451. Wannnger, L., 2012, Carrer-phase nter-frequency bases of GLONASS recevers, J Geod, 86, 139 148. Kouba J., 2009, A gude to usng Internatonal GNSS Servce IGS products, Avalable at http://gscb.jpl.nasa.gov/components/ usage.html. Unauthentcated