International Journal of Engineering Trends and Technology- Volume2Issue2- 2011 GPS User Position Using Extended Linearization Technique B.Hari Kumar #1, Dr. Dixit #2, N.Namassivaya #3, G.V.Chalapathi Rao #4 Professor & Head, ECE Department, M.V.S.R.Engineering College, Hyderabad, A.P., India. Professor, ECE Department, Allahabad University, Allahabad, U.P., India Associate Professor, ECE Department, M.V.S.R.Engineering College, Hyderabad, A.P., India Assistant Professor, ECE Department, M.V.S.R.Engineering College, Hyderabad, A.P., India Abstract - The position of a GPS receiver can be determined by are observed. GPS data of Chitrakut station in RINEX format has obtaining pseudoranges from a minimum of four different GPS been used for this purpose. satellites. The measured ranges do not represent the true ranges as the signal coming from a GPS satellite will be contaminated by Keywords - Earth Centered Earth Fixed (ECEF), Global various errors like Ephemeris error; Propagation error in the form Positioning of Ionospheric and Tropospheric delays; Satellite and Receiver Independent Exchange (RINEX). System (GPS), GPS Time (GPST), Receiver clock biases with respect to GPS Time (GPST); Multipath error etc. Most of these errors can be estimated accurately and can be I. INTRODUCTION accounted for. After making necessary corrections to the observed The Global Positioning System is all weather, space based pseudoranges the receiver position in ECEF coordinates (xu, yu, zu) navigation system. It is a constellation of a minimum of 24 and the receiver clock bias with GPS Time (GPST) tu can be determined either by using Linearization Technique [1] or Method of least squares using Bancroft Algorithm [2]. As it is easy to implement, most of the GPS receivers employ former method for satellites in near circular orbits, positioned at an approximate height of 20,150 km. above the earth. The satellites travel with a velocity of 3.9 km/sec with an orbital period of 11 hours 58 fixing the user position when a great degree of accuracy is not minutes. The satellites transmit C/A (Coarse/Acquisition) and required. The latter method is found to be more accurate when P (Precision) codes, Navigation message, Clock parameters etc. pseudoranges from more than four satellites are considered [3]. So at far only Bancroft method has been suggested for an accurate L2 (1227.60 MHz). These signals can be received and processed estimation of a GPS receiver position when more than four to obtain the user position in 3-D (latitude, longitude and satellites are observed. In this paper we are proposing a new altitude), velocity and time accurately. technique namely, Extended Linearization Technique (ELT) for the estimation of an user position. Results show that the accuracy obtained using ELT is better compared to the Linearization Technique and is on par with the Bancroft algorithm for over specified cases when pseudoranges from more than four satellites ISSN: 2231-5381 two frequencies namely L1 (1575.42 MHz) and The user estimates an apparent or pseudorange to each SV (Satellite Vehicle) by measuring the transit time of the signal. The measured ranges do not represent the true ranges as the signal coming from a GPS satellite will be contaminated by http://www.internationaljournalssrg.org Page 34 International Journal of Engineering Trends and Technology- Volume2Issue2- 2011 various errors. Most of these errors can be estimated accurately technique makes use of only four observables at a given epoch and can be accounted for. After making necessary corrections to time, this proposed technique makes use of all the observations the observed pseudoranges the receiver position in ECEF thereby accuracy in determining the position of a GPS user is coordinates (xu, yu, zu) and the receiver clock bias with GPS improved. To determine the user position in three dimensions Time (GPST) tu can then be determined. If the unknown (xu, yu, zu) and the receiver clock offset tu, pseudoranges are to coordinates of the user position are represented by xu, yu and be obtained from a minimum of four satellites. zu and the known positions of Satellite Vehicles are with xj, yj, zj, ( where j = 1,2,3,4) in ECEF coordinate system, the user ρj x x u y j y u z j z u ct u ; 2 j 2 j = 1,2,3……,m. position (in 3-D) and time offset ‘tu‘ are obtained by simultaneously solving the nonlinear equations given below. ρj x x u y j y u z j z u ct u ; 2 j 2 j = 1,2,3,4. 2 (1) Where ‘c’ is the free space velocity of electromagnetic wave in m/s. 2 (2) Where m is the number of satellites observed. The resulting equations can be written as a function of user coordinates and clock offset as j f xu , y u , z u , t u (3) Using an approximate position location ( x̂ u , ŷ u , ẑ u ) and time The measured ranges do not represent true ranges as the signal bias estimate t̂ u , an approximate pseudorange can be calculated coming from a satellite is affected by various errors like ˆ j x j xˆ u 2 y j yˆ u 2 z j zˆ u 2 c tˆu ; ephemeris error, propagation error in the form of ionospheric j =1, 2, 3, -------, m and tropospheric delays, satellite and receiver clock biases with respect to GPST, multipath error etc. In order to determine the (4) = f ( xˆ u , yˆ u , zˆ u , tˆu ) receiver position accurately, all these errors have to be estimated The unknown user position and receiver clock offset are and compensated for. In this paper, the ionospheric delay is considered to consist of an approximate component and an estimated using Klobuchar model [4]. Hopfield model has been incremental component as stated below. used for the estimation of tropospheric delay [5]. Satellite x u xˆ u x u clock bias and the relativistic effects also have been estimated y u yˆ u y u and accounted for. Finally the user position is estimated using the Linearization technique, Method of least squares using Bancroft algorithm and also by the proposed Linearization Technique. The results show that ELT is more accurate in determining the user position than the linearization method and is comparable to Bancroft algorithm. II. EXTENDED LINEARIZATION TECHNIQUE (ELT) z u zˆ u z u t u tˆu t u This allows writing Eq. (3) as f xu , yu , zu ,tu f xˆu xu , yˆu yu , zˆu zu ,tˆu tu This latter function can be expanded about the approximate point using a Taylor series. It can be shown that If more than four GPS satellites are observed at a given epoch time, using this proposed Extended Linearization Technique a (5) ˆ j j x j xˆ u rˆj xu y j yˆ u rˆj yu z j zˆu rˆj zu ct u (6) better accuracy can be obtained. Where as the Linearization ISSN: 2231-5381 http://www.internationaljournalssrg.org Page 35 International Journal of Engineering Trends and Technology- Volume2Issue2- 2011 x Where rˆj This procedure is to be repeated for about 50 times until the xˆu y j yˆ u z j zˆu 2 j 2 2 required accuracy is obtained. j ˆ j j a xj x j xˆ u rˆj ; a yj (7) y j yˆ u rˆj ; a zj III. RESULTS AND CONCLUSIONS RINEX data from Chitrakut station (Near IIT Kanpur) is used z j zˆ u (8) rˆj for this purpose [7]. The observation data of 3rd January 2006 at Where axj, ayj and azj terms denote the direction cosines of the (0 hrs. unit vector pointing from the approximate user position to the jth (SV PRN. Nos. 3 13 16 19 20 23 27) are observed at the satellite. epoch time. Algorithms have been implemented to sort out the Rewriting Equation (6) results into ephemeris data into matrix format and for the determination of j a xj x u a yj y u a zj z u ct u (9) 0 min. 30 sec.) have been used. Seven satellites satellites’ position at the epoch time [6]. By using clock correction parameters which are available as part of the When pseudorange measurements are made to ‘m’ satellites (j = m), Equation (9) can be represented in matrix form as 1 2 3 will be m x 1 matrix; ....... m x u y u x = z u ct u Navigation message, the satellite clock bias and error due to relativistic effect have been obtained. The Ionospheric delay has been estimated using Kloubachar model. All the eight coefficients for the implementation of Kloubachar model are available as part of Navigation message. The Tropospheric delay has been estimated using Hopfield method. The estimated errors and the corrected ranges have been represented in Table 1. The receiver position is then determined using the Linearization technique, Bancroft algorithm and also by the will be 4 x 1 matrix and proposed ELT. All the calculations have been carried out by writing programs in MATLAB. The results are summarized below. The (x, y, z) positions in meters of the seven observed a x1 a y1 a z1 1 a x 2 a y 2 a z 2 1 H a x 3 a y 3 a z 3 1 will be m x 4 matrix. ..... ...... ...... ..... 1 a xm a ym a zm satellites with PRN nos. of 3 13 16 19 20 23 27 at 0 hours, 0 minutes and 30 seconds of 3rd January 2006 are found to be 1.0e+007 * [-1.14435581932368 2.18537228998174 0.92840515634504 0.88498653721608 1.52115049991917 1.98379922835602 We can obtain error matrix x from the following equation x H 1 (10) As H is not a square matrix its inverse can be obtained using Inv (H) = ( Inv ( H * H ) * H ) (11) -1.28799462471086 0.84279115293681 2.17295977908060 -0.62238562333828 2.55024173739922 -0.38396284978272 1.04260459627803 2.18281560737286 -1.10756652807472 0.15114376130666 2.36981504953570 1.16498729017268 1.98311575365209 0.65606228041700 1.72062794938024] where H is the transpose of H. ISSN: 2231-5381 http://www.internationaljournalssrg.org Page 36 International Journal of Engineering Trends and Technology- Volume2Issue2- 2011 TABLE 1 ESTIMATION OF GPS ERRORS AND CORRECTION OF PSEUDORANGES Sv. no Azimu th (deg) Elevati on (deg) Observed Pseudoranges (m) Sv. clock+ relativistic (m) Iono delay (m) Tropo Delay (m) Corrected Pseudoranges (m) 3 89.75 46.29 21345372.96948 19048.06 1.9858 3.31 21364414.7719640 13 315 53.02 21123433.31848 9807.00 1.8118 2.996 21133235.1936572 16 45 21.03 23647148.85446 6064.55 3.1902 6.632 23653202.7278275 19 135 39.19 22030908.95548 -7308.55 2.2292 3.785 22023593.4330683 20 180 25.42 23234206.55447 -10893.93 2.9077 5.555 23223303.5998171 23 75.96 83.30 20047262.99349 46843.87 1.5062 2.412 20094101.9563987 27 296.5 23.79 23831204.72647 8954.52 3.0088 5.909 23840149.6080318 Using the corrected pseudoranges user position is determined and [2] Bancroft. S., “An algebraic solution of the GPS equations”, the results are shown below: IEEE Transactions on Aerospace and Electronic Systems Exact User Position as per the observation data: Vol. 21 (1985) pp: 56–59. Xu = 918074.1038m, Yu= 5703773.539 and Zu =2693918.9285m. [3] B.Hari Kumar and K.Chennakesava Reddy, “Determination User position by Linearization Technique: Of GPS Receiver Position Using Linearization Technique Xu= 918050.65m, Yu= 5703751.91m and Zu = 2693899.70m. And Bancroft Algorithm”, IETECH User position by Bancroft Algorithm: Communication Techniques, Volume-2, Xu = 918075.38m, Yu = 5703776.40m and Zu = 2693918.73m. pp-14-16, 2008 User position using Extended Linearization Technique Journal of Number-1, [4] Klobuchar J, “ Design and characteristics of the GPS Xu = 918075.72 m, Yu = 5703777.12 m and Zu = 2693918.91 m ionospheric time – delay algorithm for single frequency users”, Proceedings of PLANS’86 – Position Location and Results show that the proposed method namely the Extended Navigation Symposium, Las Vegas, Nevada, November Linearization 4-7, pp: 280-286. Technique is more accurate compared to Linearization Technique and is comparable to Bancroft algorithm [5] Hopfield HS, “Two – quartic tropospheric refractivity in determination of the user position when data from more than profile for correcting satellite data”, Journal of Geophysical four satellites are taken into account. research, Vol. 74, No. 18, pp: 4487-4499, 1969. [6] Strang, G. and Borre, K., “Linear Algebra, Geodesy, and REFERENCES [1] Kaplan, E. D., "Understanding GPS: Principles and GPS”,Wellesley-Cambridge, Wellesley, MA, 1997. [7] http://home.iitk.ac.in/~ramesh/gps/gpsdata/gpsdata.html Applications", Artech House, 1996. ISSN: 2231-5381 http://www.internationaljournalssrg.org Page 37