IRNSS/NavIC Single-Point Positioning: A Service Area Precision Analysis

Transcription

1 This is an Autho's Oiginal Manuscipt of an aticle published by Taylo & Fancis in Maine Geodesy on 12/12/2016 available online at IRNSS/NavIC SinglePoint Positioning: A Sevice Aea Pecision Analysis Safooa Zaminpadaz *, Pete J.G. Teunissen, and Nandakumaan Nadaajah GNSS Reseach Cente, Cutin Univesity, Peth, Austalia *: safooa.zaminpadaz@postgad.cutin.edu.au The Indian Regional Navigation Satellite System (IRNSS) has ecently (May 2016) become opeational. The system has been developed with the objective of offeing positioning, navigation and timing (PNT) to the uses in its two sevice aeas, coveing the Indian landmass and the Indian Ocean espectively. It is the goal of this contibution to povide futhe insight into the fullconstellation L5 pseudoange singlepoint positioning (SPP) capabilities of the system. A detailed dilution of pecision (DOP) analysis of its two sevice aeas, including the identification, in location and time, of poo eceivesatellite geometies is povided. It is heeby demonstated how the impact of some of these poo eceivesatellite geometies can be mitigated by means of heightconstaining. An oveview and analysis of the SPP pecision is also povided including easytouse epesentative dayaveaged values fo a gid of locations coveing the two sevice aeas. Keywods: Indian Regional Navigation Satellite System (IRNSS); Navigation with Indian Constellation (NavIC); Dilution of Pecision (DOP); L5 Pseudoange; SinglePoint Positioning (SPP) 1 Intoduction The Indian Regional Navigation Satellite System (IRNSS) has ecently (May 2016) become opeational and povided with the opeational name of NavIC (Navigation with Indian Constellation). It has been developed by the Indian Space Reseach Oganization (ISRO) with the objective of offeing positioning, navigation and timing (PNT) to the uses in its sevice aea. It theeby povides two types of sevices: a standad positioning sevice (SPS) fo civilian uses and a esticted sevice (RS) fo authoized uses (ISRO 2014). As IRNSS has ecently become fully opeational, it is impotant to gain an undestanding of its navigational potential and actual positioning capabilities fo civilian uses. This is not only of geneal inteest, but also specifically fo uses opeating in its sevice aea. Thee ae to date only a few published studies on IRNSS s positioning pefomance and none wee based on the full constellation. Among the published IRNSS studies, seveal ae simulationbased (Mozo Gacia et al. 2010, Sama et al. 2010, Seka et al. 2012, Rethika et al. 2013, Rao 2013, Su et al. 2012), while some ae based on eal data, such as Thoelet et al. (2014) in which the clock stability of IRNSS1A is assessed. The accuacy of a pecise model fo sola adiation pessue is tested using the IRNSS1A and 1B obsevations in (Kumai et al. 2015). Babu et al. (2015) compaes obit detemination methods fo IRNSS1A, 1B and 1C. In ode to validate the obit accuacy with modenized ephemeis paametes, Chandasekha et al. (2015) employs the IRNSS1A, 1B and 1C eal data. Montenbuck and Steigenbege (2015) used the obsevations of the IRNSS1A and 1B to investigate the quality of the IRNSS navigation messages. Nadaajah et al. (2015), afte assessing the IRNSS noise chaacteistics, combines the L5/E5 signals of IRNSS, GPS, 1

2 Galileo and QZSS fo instantaneous attitude detemination. Some positioning esults ove India based on the 4satellite data of I1, I2, I3, and I4 ae pesented in (Ganeshan et al. 2015, Pal and Ganeshan 2015), and the position accuacy of two IRNSS satellites integated with the othe satellite systems is pesented in (Thombe et al. 2016). Zaminpadaz et al pesents the fist IRNSS standalone positioning esults ove Austalia and Odijk et al. (2016) pesents the fist analysis of the diffeential intesystem biases (DISBs) between L5 signal of IRNSS w..t. the L5/E5a signals of GPS, Galileo and QZSS. The goals of this contibution ae to povide insight into IRNSS and to demonstate its fullconstellation singlepoint positioning (SPP) pefomance. As such this contibution povides the fist IRNSS standalone positioning pefomance analysis fo diffeent locations in its two sevice aeas. This contibution is oganized as follows. In section 2 the full IRNSS constellation with its two sevice aeas ae descibed as well as the SPP model that foms the basis of ou analyses. Section 3 povides a detailed dilution of pecision (DOP) analysis of IRNSS s two sevice aeas, including the identification, in location and time, of poo eceivesatellite geometies. It is heeby also demonstated how the impact of some of these poo eceivesatellite geometies can be mitigated by means of heightconstaining. An oveview and analysis of the SPP pecision fo a gid of locations ove the two sevice aeas is given in section 4. It includes easytouse epesentative dayaveaged values of the positioning pecision. 2 The IRNSS/NavIC System In this section we descibe the IRNSS constellation with its two sevice aeas as well as the singlepoint positioning (SPP) model that foms the basis of ou analyses. The IRNSS Constellation The IRNSS constellation consists of thee geostationay obit (GEO) satellites and fou inclined geosynchonous obit (IGSO) satellites (see Figue 1 and Table 1). The obital peiod of the IRNSS satellites is one sideeal day (23 hous and 56 minutes), such that the IRNSS satellite gound tacks epeat evey sola day (24 hous) fou minutes ealie. The IRNSS satellites tansmit navigation signals, based on Code Division Multiple Access (CDMA), on L5 ( MHz) and on S ( MHz), with a Binay PhaseShift Key (BPSK (1)) modulation fo standad positioning sevice (SPS) uses, and with a Binay Offset Caie (BOC (5,2)) modulation fo esticted sevice (RS) uses (ISRO 2014). The IRNSS is established with the objective of offeing positioning, navigation and timing (PNT) sevices to the uses in its sevice aea. The IRNSS classifies its sevice aeas boadly into the two egions shown in Figue 2. The pimay sevice aea of IRNSS encompasses the Indian landmass and a egion lying within a distance of 1500 km fom its geopolitical bounday, and the seconday sevice aea extends between latitudes 30 S to 50 N and longitudes 30 E to 130 E (Sama et al. 2010, Ganeshan 2012, Saikian and Vikam 2013). Also the seconday sevice aea is impotant as it coves the Indian Ocean being one of the busiest oceans. The Indian Ocean is the thid lagest of the wold s oceanic divisions, coveing appoximately 20% of the wate on the Eath s suface. It contains majo sea outes connecting the Middle East, Afica, and East Asia with Euope and the Ameicas, and it povides an estimated 40% of the wold s offshoe oil poduction. 2

3 The SingleEpoch SPP Model We assume that a single eceive is tacking m IRNSS satellites on fequency L5. The singleepoch, singlefequency lineaized SPP obsevation equations can then be fomulated in vectomatix fom as x E{ p} = [ G em ] dt (1) D{ p } = 2 W 1 whee E {.} and D {.} denote the expectation and dispesion opeatos, espectively. The m vecto p contains the obsevedminuscomputed pseudoange obsevables to the m IRNSS satellites. They ae a pioi coected fo the ionospheic delays and fo the dy topospheic delays. The m 3 matix G = [ ] 1 m T u,, u contains the unit diection vectos s u, s = 1,, m, fom the eceive antenna to the m tacked IRNSS satellites and e m is the m vecto of ones. The 3 vecto x is the eceive position incement, and the scala dt is the incement of the eceive clock eo biased by the eceive code hadwae delay. The paamete denotes the zenithefeenced use ange accuacy which captues the obsevables noise chaacteistics as well as the emaining unmodeled effects. The mm diagonal weight matix W = diag ( w 1,, m w ) contains the elevationdependent weights (Eule and Goad 1991), s w s = [1 10exp( )] 10 2 (2) s with being the elevation of the satellite s with espect to eceive. Assuming that the design matix G, e ] is of full ank, the SPP weighted leastsquaes solution of the position vecto is given as [ m xˆ Q xˆ xˆ = = ( G W T 2 D T G ( G W D ) 1 G G W p ) T 1 D (3) in which W D = D ( D W m T m 1 D m ) 1 D T m and the ange space of the m ( m 1) diffeence matix T D m spans the othogonal complement of e m, i.e. D mem = 0. By making use of the matix T 1 T identity WD = W ( Im em ( emw em ) em W ), the vaiance matix of xˆ can also be expessed as m 2 s s s T 1 xˆ ˆ ( [ ][ ] ) x s1 Q w u u u u (4) 3

4 with the weighted aveage m s s m s / s1 s1 u w u w. This expession now explicitly shows how the eceivesatellite unit diection vectos and thei coesponding satellite elevation weighting contibute to the vaiance matix. HeightConstained Model As we will see in the following, thee ae situations fo which the IRNSS SPP model tuns out to be too weak to povide accuate positioning. Heightconstaining of the model would then be a possible emedy. Infomation on the height component could be povided if the vetical use position would be known on aveage and not vay too much. Examples of such applications ae maine positioning, bathymetic suveying, and kinematic positioning ove small aeas with low height fluctuations (Zhu and Santee 2002; Godha and Cannon 2007). When enfocing the weighted height constaint, the obsevational model in (1) is extended with 2 E h} = [0, 0, 1] x, D{ h} = (5) { h in which h denotes the height constaint coected fo an initial height value and h is the a pioi standad deviation of the height constaint. 3 IRNSS PDOPHDOP Analysis In this section we povide a singlepoint positioning dilution of pecision analysis fo a gid of locations coveing IRNSS s two sevice aeas. Hoizontal Dilution of Pecision We use the DOP as a scala measue to diagnose the stength of the eceivesatellite geomety. The Position DOP (PDOP) is defined as (Teunissen 1998, HofmannWellenhof et al. 2013) T PDOP = tace [ G W G ] tace( Q ) (6) 5 D The PDOP consides the impact of the eceivesatellite geomety on all thee eceive coodinates. The Hoizontal DOP (HDOP) is obtained in case only the hoizontal eceive coodinates ae consideed. In the following we povide the 24hou PDOP and HDOP time seies fo diffeent locations ove the two sevice aeas. As the IRNSS eceivesatellite geomety epeats itself evey day fou minutes ealie w..t the pevious day, the signatues of these time seies ae epesentable fo any day. Figue 2 shows the selected locations, fo the pimay sevice aea in ed and fo the seconday sevice aea in black. xˆ xˆ 4

5 Pimay Sevice HDOP Figue 3 shows the time seies of the numbe of visible satellites (black), the PDOP (ed) and the HDOP (blue) fo the pimay sevice aea locations (see Figue 2). It can be seen that the PDOP and HDOP values ae easonably constant ove the day fo the locations at latitudes of 0 and 20, with PDOP 4 and HDOP 2. Fo locations (φ = 40, λ = 65 ) and (φ = 40, λ = 101 ) we note a diffeence in satellite visibility, despite thei symmety w..t. the location of satellite I3. This is due to the fact that the distibution of the othe fou IGSOs and two GEOs is not symmetic w..t the I3 location. Fo example, I6 is located 50.5 to the West of I3 wheeas I7 is located 46.5 to the East of I3. Theefoe, location (φ = 40, λ = 65 ) has a longe visibility of I7 than location (φ = 40, λ = 101 ) has of I6. Still, except fo a distinct peak in the PDOP time seies of location (φ = 40, λ = 101 ) ove the peiod UTC [17:32:0019:23:00], the PDOP and HDOP values ae compaable to those of locations (φ = 40, λ = 65 ) and (φ = 40, λ = 83 ), with PDOP 6 and HDOP 4. The occuence of the PDOP peak at location (φ = 40, λ = 101 ) can be explained though the coesponding eceivesatellite geomety. Figue 4 depicts the satellites skyplot at UTC 18:30:30, which is the moment the PDOP peak occus at location (φ = 40, λ = 101 ). It can be seen that at this moment in time the eceivesatellite lineofsight unit vectos of all five satellites I1, I3, I4, I5 and I7 fom a conelike geomety (Teunissen 1990, Zaminpadaz et al. 2016), thus leading to a nea ankdefect SPP design matix and hence lage PDOP values. Note that the pooest estimable diection, i.e. the diection of the cone symmety axis, has almost no component in the hoizontal plane. This explains why the lage peak in the PDOP at location (φ = 40, λ = 101 ) (see Figue 3) is absent fom the HDOP. Seconday Sevice HDOP Figue 5 shows the time seies of the numbe of visible satellites (black), the PDOP (ed) and the HDOP (blue) fo the seconday sevice aea locations (see Figue 2). When we conside the behavio of the DOPvalues, fou diffeent categoies of locations can be disciminated: I, II, III and IV. Fo the locations in categoy I, the DOPvalues ae almost constant and simila to those of the pimay sevice aea, PDOP 4 and HDOP 2.5. Also the DOPvalues of the locations within categoy II ae easonably constant, although somewhat moe iegula than those of I. The fist time that we eally see a PDOPpeak occuing is at the locations of categoy III, the explanation of which is the same as given fo pimay sevice aea location (φ = 40, λ = 101 ). Moe fequent and moe exteme DOPvalues, both fo PDOP and HDOP, occu when we go to the bode of the seconday sevice aea at locations inside categoy IV. At these locations pope positioning is eally becoming toublesome. The lage DOPvalues at the bode locations can be mitigated by imposing a height constaint (cf. 5). In Figue 6, the heightconstained ( h / 1), pactically coinciding PDOP and HDOP time seies ae pesented fo bode locations in categoy IV. It can be seen that upon constaining the height component, nonequato locations indeed obtain a easonably smooth DOP behavio fo the whole 24hou peiod. Howeve, fo the equato locations, thee still exist two time intevals ove which the heightconstained DOP eaches 5

6 lage values. This is due to the fact that at these two locations only fou satellites, two GEOs and two IGSOs, ae visible and theefoe thee ae two peiods when the two IGSOs occupy (almost) the same skyplot position, thus leading to a poo eceivesatellite geomety with coesponding lage DOP values. DayAveaged HDOP Map With the above given insight into the daily PDOP and HDOP behavio of the selected locations in the pimay and seconday sevice aeas, we now pesent a moe fluid epesentation of the spatial HDOP vaiability. Figue 7 shows the 24hou aveage IRNSS HDOP colo map ove its sevice aeas. As the IRNSS eceivesatellite geomety epeats itself appoximately evey day, this aveage HDOP map can be consideed a useful dayaveage. When using this value as an appoximation to the actual HDOP, one should keep in mind though that this appoximation will be pooe fo locations that show a geate HDOP timevaiability, such as when one gets close to the bodes of the seconday sevice aea. The map can be used to compute the aveage hoizontal positioning standad deviation fo a paticula location l as l HDOPl (7) 2 in which HDOPl denotes the aveage HDOP value of that location. 4 IRNSS SPP Pecision In this section an oveview is povided of the expected IRNSS s singlepoint positioning pecision ove the two sevice aeas. SPP Scatteplots and Confidence Ellipses To get a fist impession of the SPP epeatability, we have pocessed five days of 30second IRNSS data collected with the cutoff elevation angle of 10 at two diffeent stations in Westen Austalia. The two stations ae Peth (φ = 32, λ = ) and Fitzoy (φ = 18.13, λ = ) which ae equipped with JAVAD TRE_G3TH_8 eceive and SEPTENTRIO POLARX5 eceive, espectively. Peth is close to location (φ = 30, λ = ) of the seconday sevice aea, while Fitzoy is east of location (φ = 15, λ = ) and close to the bode of the seconday sevice aea. The data ae a pioi coected fo the ionospheic delays using the pedicted Global Ionosphee Map GIM (CODE 2016) and fo the dy topospheic delays using the Saastamoinen model (Saastamoinen 1972). Thei use ange accuacies wee detemined as 1.80 m fo Peth and 2.30 m fo Fitzoy by means of vaiance component estimation (VCE) (Teunissen and AmiiSimkooei, 2008). Fo the elatively cude pseudoangeonly SPP model, these estimates will include all emaining unmodelled effects, such as e.g., satellite obital eos, satellite clock eos and esidual atmospheic delays. That of Fitzoy is somewhat lage due to the additional ionospheic uncetainty. The singleepoch SPP scatteplots of the two stations ae shown in Figue 8 (a 6

7 and c). Supeimposed on the scatteplots ae also shown the fomal (geen) and empiical (ed) 95% confidence ellipses, which show a easonable good ageement. But note that the two SPP scatteplots ae not as homogeneous as one would expect if data would be coming fom a nomal distibution. The eason fo this lack of homogeneity lies in the pesence of mismodelled effects, paticulaly due to the satellite clocks and ionospheic delays. These effects would be absent if one would conside elative point positioning (RPP) between two neaby stations. This is demonstated in Figue 8 (b) which shows the RPP scatteplot of Peth and a closeby station. Note that the Fitzoy scatteplot is fa moe elongated than the one of Peth (Figue 8, a and c). This is due to the elatively poo eceivesatellite geomety of Fitzoy, being so close to the bode of the seconday sevice aea. Consideable impovement is possible though if one applies a height constaint. Figue 8 (d) shows the Fitzoy scatteplot when a height constaint of h 1m is applied. Also note that all confidence ellipses of Figue 8 ae oiented in a nothwestely diection. This can be explained by means of the eceivesatellite geomety and its impact on the confidence ellipse of x ˆ, ( xˆ x ) Q ( xˆ x ) (8) T 1 2 xˆˆ x 2 in which the constant is chosen such that a cetain confidence level is eached. As the 1 diection of elongation is given by the diection of the eigenvecto of coesponding to Q xx ˆˆ its smallest eigenvalue, it follows with the aid of (4) that this diection is given by T f ag min f Q f f m s1 1 xx ˆˆ s T s ag min w [ f ( u u )] f 2 (9) Figue 9 depicts the dayaveaged skyplot position of the IRNSS satellites as well as that of the weightedaveage u at Peth on DOY 153 of 2016 with cutoff angle of 10. As the s diffeences ( u u) ae mainly oiented along the NothEast diection, the diection f that minimizes thei contibution to (9) will mainly lie in a NothWestely diection. Pimay sevice aea We now povide an oveview of the SPP pecision fo the pimay sevice aea. Table 2 gives fo the pimay sevice aea locations (within the gay box), the dayaveaged values of thei Noth and East standad deviations, and, and coesponding coelation coefficients NE N. Fo each location, thee ae two ows of values; the fist ow coesponds to the unconstained scenaio with 1m while the second ow to the heightconstained scenaio with 1m and / 1. Given the use ange accuacy of 1m, the ate of change of the h E 7

8 hoizontal pecision w..t. the vaiation of the atio h / can be descibed consideing two cases h / 1and h / 1. The aveage ate of change of the hoizontal pecision w..t. h / 1 is zeo fo the locations within (30 <φ < 40, 65 <λ < 110 ), 0.01, 0.25, 0.14, 0.43 and 0.22 fo the locations within (λ = 50.5, ), (φ < 40, λ = 30 ), (φ < 40, λ = 130 ), (φ = 40, λ = 30 ) and (φ = 40, λ = 130 ), espectively. The aveage ate of change of the hoizontal pecision w..t. h / 1is 0.07, 0.03, 0.02, 0.02, 0.02, 0.04 and 0.06 fo the locations on (λ = 30, 50.5, 65, 83, 110, and 130 ), espectively. Fo any othe values of use ange accuacy, one can scale these ates of change and the esults in Table 2 accodingly. Fo example, fo 3m, these values should be multiplied by 3. The tablevalues can be consideed epesentative fo the whole day as thei HDOP time seies showed a athe stable behavio ove the 24hou peiod, see Figue 3. It follows fom Table 2 that location (φ = 0, λ = 83 ), being the gound tacks symmety point, has the best pecision and that the pecision gets pooe in noth/south diection the futhe one depats fom the equato and likewise pooe in east/west diection the futhe one moves away fom this location in such east/west diection. As the impact of heightconstaining is lage the pooe the unconstained pecision, its effect is felt moe the futhe one is fom location (φ = 0, λ = 83 ). But oveall the effect of heightconstaining on these pimay sevice aea locations is only maginal. Seconday Sevice aea We now povide an oveview of the SPP pecision fo the seconday sevice aea. Table 2, outside the gay box, gives an oveview fo the seconday sevice aea locations. All the dayaveaged enties of the table can be consideed epesentative fo the whole day as thei coesponding HDOP time seies ae shown to be sufficiently stable, see Figue 5. Note that the fomal esults given fo location (φ = 30, λ = ), which is close to Peth, is indeed consistent with the empiical esults obtained fo Peth (Figue 8 a). As was mentioned fo the pimay sevice aea locations, the esults of the seconday sevice aea show the geneal tendency that the SPP pecision gets pooe the futhe one moves away fom the cental location (φ = 0, λ = 83 ). Howeve, fo the locations on the bode of the seconday sevice aea, with longitudes of λ = 30, 130, the wosening in the SPP pecision is such that no epesentative dayaveage can be given. As thei unconstained HDOPvalues show vey lage excusions ove the day (cf. Figue 5), no dayaveaged values have been given fo these locations in Table 2. Fo thei heightconstained case the situation is diffeent. Upon constaining the height component, the HDOP time seies at locations with latitudes of φ = 30, 30, 40 almost flatten and the coesponding dayaveaged pecision can be consideed epesentative again (cf. Figue 6). This is howeve not the case fo the equato and neaequato locations with latitudes of φ = 15, 0, 20. Thei heightconstained dayaveaged values ae theefoe excluded fom Table 2. To have a bette desciption of thei pecision, we divided the 24hou peiod into two subpeiods; one subpeiod accommodating the lage peaks fo the HDOP and one containing the emaining pat of the 24hou peiod (cf. Figue 6). Note that these subpeiods ae simila fo locations with the same longitude. Table 3 lists the aveage Noth, East standad deviations and coelation coefficients fo these locations ove the subpeiods with smooth HDOP behavio. 8

9 The EastNoth coelation coefficients given in Tables 2 and 3 show that fo those locations with the same longitude o the same latitude of cental location (φ = 0, λ = 83 ) of pimay sevice aea, East and Noth components can be consideed uncoelated. Compaing the East and Noth standad deviations fo these locations, it can be said that thei coesponding hoizontal confidence ellipses ae vetically o hoizontally elongated towads location (φ = 0, λ = 83 ) of pimay sevice aea. Fo the locations with diffeent latitude and longitude fom those of location (φ = 0, λ = 83 ) of pimay sevice aea, the EastNoth coelation is such that thei coesponding confidence ellipses ae elongated towads this point. Finally we note, although the above tables povide dayaveaged values, while the IRNSS satellite gound tacks do not exactly have a 24hou epeat cycle, that the values povided can still be consideed epesentative fo any abitay day. The unconstained mean values given in Tables 2 and 3 can vay fo any othe abitay day by at most 2 cm in Noth and 4 cm in East fo the use ange accuacy of 1m. Theses vaiations get smalle when the undelying model is stengthened by a weighted height constaint. Iespective of the value of h, the lagest vaiation in the mean Noth and East standad deviations is 1 cm in case 1m. 5 Summay and Conclusions As IRNSS has ecently (May 2016) become opeational, it is impotant to gain an undestanding of its navigational potential and actual positioning capabilities fo civilian uses. This contibution has povided fo that pupose a fist full constellation IRNSS singlepoint positioning pecision analysis ove its two sevice aeas. A dilution of pecision analysis of daily PDOP and HDOP time seies was given, including the identification, in location and time, of poo eceivesatellite geometies. The natue of these geometies was explained and it was demonstated how the impact of some of these poo eceivesatellite geometies can be mitigated by means of heightconstaining. We also povided an oveview and analysis of the singlepoint positioning pecision fo a gid of locations coveing the two sevice aeas. It includes, with the exception of six locations, easytouse epesentative dayaveaged values of the positioning pecision, unconstained as well as heightconstained. Fo the six exempted locations, on and nea the equato on the bode of the seconday sevice aea, subday aveages wee given. Acknowledgements The second autho is the ecipient of an Austalian Reseach Council (ARC) Fedeation Fellowship (poject numbe FF ). A pat of the IRNSS date was povided by Geoscience Austalia. This suppot is gatefully acknowledged. 9

10 Refeences Babu R, Mula P, Ratnakaa SC, Ganeshan AS IRNSS satellite paamete estimation using combination stategy. Global Jounal of Science Fontie Reseach, 15(3): Chandasekha MV, Rajaajan D, Satyanaayana G, Timal N, Rathnakaa SC, Ganeshan AS Modenized IRNSS boadcast ephemeis paametes. Contol Theoy and Infomatics, 5(2): 1 9. CODE Global ionosphee maps poduced by CODE. Accessed July ftp://ftp.unibe.ch/aiub/code/2015. Eule HJ, Goad CC On optimal filteing of GPS dual fequency obsevations without using obit infomation. Bulletin Geodesique 65(2): Ganeshan AS Oveview of GNSS and Indian Navigation Pogam. Pape pesented at GNSS Use Meeting, ISRO Satellite Cente, Bangaloe, Febuey 23. Ganeshan AS, Ratnakaa SC, Sinivasan N, Rajaam B, Timal N, Anbalagan K Fist Position Fix with IRNSS. Inside GNSS, July/August Accessed Apil Godha S, Cannon M GPS/MEMS INS integated system fo navigation in uban aeas. GPS Solutions 11(3): HofmannWellenhof B, Lichtenegge H, Collins J Global positioning system: theoy and pactice. Spinge Science & Business Media. ISRO Indian egional navigation satellite system: signal in space ICD fo standad positioning sevice, Vesion 1.0. ISRO Satellite Cente, June ISRO PSLVC33/IRNSS1G. Accessed July Kumai A, Samal K, Rajaajan D, Swami U, Babu R, Katik A, Rathnakaa SC, Ganeshan AS Pecise modeling of sola adiation pessue fo IRNSS satellite. Jounal of Natual Sciences Reseach 5(3), Montenbuck O, Steigenbege P, Riley S IRNSS obit detemination and boadcast ephemeis assessment. Pape pesented at Intenational Technical Meeting of The Institute of Navigation, Dana Point, CA, Januay pages Mozo Gacia A, Piiz R, Lainez Sampe MD, Romay Meino MM Multisystem Real Time PecisePointPositioning, today with GPS+GLONASS in the nea futue also with 10

11 QZSS, Galileo, Compass, IRNSS. Pape pesented at Intenational Symposium on GPS/GNSS, Taipei, Taiwan, Octobe Nadaajah N, Khodabandeh A, Teunissen PJG Assessing the IRNSS L5signal in combination with GPS, Galileo, and QZSS L5/E5asignals fo positioning and navigation. GPS Solutions, 20: Odijk D, Nadaajah N, Zaminpadaz S, Teunissen PJG GPS, Galileo, BDS, QZSS and IRNSS Diffeential ISBs: Estimation and Application. GPS Solutions. Accessed July doi: /s y. Pal S, Ganeshan AS Indian GNSS Paadigm. In 2015 PNT Symposium. Standfod, CA, 1213 Novembe Rao VG Poposed LOS fast TTFF signal design fo IRNSS. PhD diss. Univesity of Calgay. Rethika T, Misha S, Nimala S, Rathnakaa SC, Ganeshan AS Single fequency ionospheic eo coection using coefficients geneated fom egional ionospheic data fo IRNSS. Indian Jounal of adio & Space Physics 42(3), Saastamoinen J Atmospheic coection fo the toposphee and statosphee in adio anging satellites. In The use of atificial satellites fo geodesy, edited by S.W. Heniksen, A. Mancini and B.H. Chovitz, Washington, DC, Ameican Geophysical Union. Saikian B, Vikam V IRNSS achitectue and applications. Jounal of Comm & Electon 1(3): Sama AD, Sultana Q, Sinivas VS Augmentation of indian egional navigation satellite system to impove dilution of pecision. Jounal of Navigation 63(2): Seka SB, Sengupta S, Bandyopadhyay K Spectal compatibility of BOC (5, 2) modulation with existing GNSS signals. Pape pesented at Position Location and Navigation Symposium (PLANS), 2012 IEEE/ION, IEEE. pages Su XL, Zhan XQ, Niu MC, Zhang YH Pefomance Compaison fo Combined Navigation Satellite Systems in AsiaPacific Region. Jounal of Aeonautics, Astonautics and Aviation. Seies A 44(4): Teunissen PJG GPS op afstand bekeken. In: een halve eeuw in de goede ichting. Lustumboek Snellius , DUM Delft, Teunissen PJG A poof of Nielsen s conjectue on the elationship between dilution of pecision fo point positioning and fo elative positioning with GPS. IEEE on Aeospace and Electonic Systems 34(2): Teunissen PJG, AmiiSimkooei AR Leastsquaes vaiance component estimation. Jounal of Geodesy 82(2):

12 Thoelet S, Montenbuck O, Meue M IRNSS1A: signal and clock chaacteization of the Indian egional navigation system. GPS solutions 18(1): Thombe S, Bhuiyan MZH, Södeholm S, KikkoJaakkola M, Ruotsalainen L, Kuusniemi H A Softwae MultiGNSS Receive Implementation fo the Indian Regional Navigation Satellite System IETE Jounal of Reseach 62(2): Zaminpadaz S, Teunissen PJG, Nadaajah N IRNSS standalone positioning: fist esults in Austalia. Jounal of Spatial Science 61(1):5 27. Zhu J, Santee R Impovement of GPS phase ambiguity esolution using pio height infomation as a quasiobsevation. Geomatica 56(3):

13 Table 1 Infomation on the IRNSS satellites (ISRO 2016). Satellite Type Longitude Inclination Launch date IRNSS1A (I1) IGSO 55 E 29.0 July 2013 IRNSS1B (I2) IGSO 55 E 31.0 Apil 2014 IRNSS1C (I3) GEO 83 E Octobe 2014 IRNSS1D (I4) IGSO E 30.5 Mach 2015 IRNSS1E (I5) IGSO E 28.1 Januay 2016 IRNSS1F (I6) GEO 32.5 E Mach 2016 IRNSS1G (I7) GEO E Apil

14 Table 2 IRNSS L5 singleepoch SPP NothEast fomal standad deviations, N, E, and coelation coefficients, NE with the cutoff elevation angle of 10 ove the IRNSS sevice aea. Thee ae two ows of values fo each location; fist ow coesponds to the unconstained scenaio ( 1m) and second ow to the heightconstained scenaio ( 1m and h / 1). On each ow, thee values ae given which ae (fom left to ight) N [m], E [m], NE. These values ae the dayaveaged values fo DOY 153 of The gay box contains the esults of the pimay sevice aea. φ\λ ,4.16, ,3.00, ,3.01, ,2.49, ,1.90, ,2.16, ,1.59, ,2.01, ,1.44, ,1.93, ,1.31, ,1.97, ,1.39, ,2.15, ,1.59, ,1.52, ,1.40, ,1.29, ,1.18, ,1.18, ,1.04, ,1.12, ,0.93, ,1.15, ,0.99, ,1.29, ,1.18, ,1.21, ,1.19, ,0.98, ,0.98, ,0.86, ,0.86, ,0.78, ,0.78, ,0.82, ,0.82, ,0.98, ,0.98, ,1.67, ,1.47, ,1.42, ,1.25, ,1.32, ,1.09, ,1.26, ,0.97, ,1.29, ,1.04, ,1.42, ,1.24, ,3.13, ,2.08, ,2.71, ,1.73, ,2.44, ,1.56, ,2.31, ,1.40, ,2.37, ,1.49, ,2.71, ,1.73, ,3.53, ,2.72, ,2.71,

15 Table 3 IRNSS L5 singleepoch SPP NothEast fomal standad deviations, N, E, and coelation coefficients, NE with the cutoff elevation angle of 10 on the IRNSS seconday sevice aea bode (see Figue 3). These values ae the dayaveaged values (excluding the mentioned peiods) fo DOY 153 of 2016 fo heightconstained scenaio ( 1m and / 1). h φ λ 24hou peiod excluding two subpeiods (UTC [hh:mm]) N [m], E [m], [09:00 11:41] & [20:56 23:17] 2.47,2.48, [05:40 07:38] & [17:50 19:23] 2.41,2.30, [09:12 11:25] & [21:11 23:23] 2.15,2.23, [05:01 07:35] & [17:26 19:32] 2.05,2.12, [09:06 11:29] & [21:04 23:31] 2.31,2.36, [05:47 07:24] & [17:37 19:31] 2.29,2.21,0.23 NE 15

16 Figue 1. (Left) The 7 satellite IRNSS constellation. (Right) The 3 GEO and 4 IGSO satellite gound tacks of IRNSS. 16

17 Figue 2. Pimay (ed) and seconday (black) sevice aea locations. The inne and oute geen boundaies indicate the bode of the pimay and seconday sevice aeas, espectively. 17

18 Figue 3. Pimay sevice aea time seies of the numbe of visible IRNSS satellites (black) on DOY 153 of 2016, with 10 cutoff angle, and the coesponding PDOP (ed) and HDOP (blue) values. 18

19 Figue 4. IRNSS satellites skyplot at UTC 18:30:30 of pimay sevice aea location (φ = 40, λ = 101 ) on DOY 153 of 2016, with 10 cutoff angle. The cicles and the staight lines ae the axes fo the elevation and azimuth, espectively. The cone symmety axis (pooest estimable diection) is identified by the puple dot and the cicula contou lines show the loci of the diections that make angles of 47 and 49 with this symmety axis. 19

20 Figue 5. Seconday sevice aea time seies of the numbe of visible IRNSS satellites (black) on DOY 153 of 2016, with 10 cutoff angle, and the coesponding PDOP (ed) and HDOP (blue) values. 20

21 Figue 6. Time seies of the numbe of visible IRNSS satellites and the coesponding heightconstained PDOP (ed) and HDOP (blue) with h / 1ove the seconday sevice aea, on DOY 153 of 2016 with 10 cutoff angle. 21

22 Figue 7. Dayaveaged IRNSS HDOP colo map fo DOY 153 of 2016 with 10 cutoff angle. 22

23 Figue 8. Singleepoch (30sec sampling ate) scatteplots fo five DOYs of 2016 with 10 cutoff angle. Fomal (geen) and empiical (ed) 95% confidence ellipses ae also shown. (a) Peth SPP scatteplot (close to location (φ = 30, λ = ) of seconday sevice aea); (b) Peth RPP scatteplot; (c) Fitzoy SPP scatteplot (east of location (φ = 15, λ = ) close to bode of seconday sevice aea); (d) Fitzoy heightconstained SPP scatteplot ( 1m). h 23

24 Figue 9. Dayaveaged IRNSS skyplot at Peth fo DOY 153 of 2016 with 10 cutoff angle. The cicles and the staight lines ae the axes fo the elevation and azimuth, espectively. The ed cicle indicates the skyplot position of vecto u. 24