Method for locating a vehicle on a predefined trajectory, computer program and associated device

Abstract

The invention relates to a method for locating a train (4) moving on a set of tracks (V1, V2) on which virtual beacons defined by their geographical coordinates are distributed and equipped with an on-board satellite receiver receiving geopositioning signals from satellites S1, …; Ss i1 / for k = 1 to s: calculation of the correlation delay, between the geostationary signal received by the satellite receiver at time t from satellite Sk and the theoretical geostationary signal calculated as having to be received from satellite Sk at the position of the virtual beacon Bi; i2 / for the beacon considered Bi: calculation of representative values ​​of differences between some of said delays calculated for k = 1 to s and at least one value determined as a function of one or more other(s) of said calculated correlation delays;i3 / estimation of the train's location based on the aforementioned representative values ​​of the differences between the calculated delays. Figure for the abbreviation: Fig. 1;

Description

Title of the invention: Method for localizing a vehicle on a predefined trajectory, computer program and associated device. Technical field

[0001] The present invention relates to the field of position estimation carried out by satellite geolocation, by a satellite positioning system also designated by the acronym GNSS (for Geolocation and Navigation by a Satellite System). Previous technique

[0002] Two categories of GNSS positioning are distinguished: absolute positioning, where location is determined solely by GNSS pseudo-distances without any external information, and differential or relative positioning, where positioning is determined relative to a reference (ground reference stations). The latter includes solutions such as PPP (Precise Point Positioning), RTK (Real-Time Kinematic Positioning), or SB AS (Satellite Complement to a GNSS System).

[0003] To achieve the required Safety Integrity Level performance levels, SIL4, a ground infrastructure is currently required (including actual beacons).

[0004] The applicant has previously developed a positioning solution based on a correlation calculation between a received GNSS signal and a predicted PRN code for a known position on georeferenced virtual beacons: without errors, a maximum correlation power should be observed for the assumption of the position closest to the receiver, allowing the detection of the possible positions of the virtual beacons closest to the receiver.

[0005] The absolute accuracy of a position obtained by a position determination system, traditionally by triangulation of data from satellites and received by a satellite receiver, is subject to errors which can be expressed and summed as follows:

[0006] UERE2 = (E_SIS)2 + (E_Tropo) 2+ (E_Iono)2 +(E_Rx)2 +(E_Multitrajet)2 where

[0007] the UERE (“User Equivalent Range Error”) is the user equivalent range error;

[0008] E_sis represents the errors of the "Signal In Space" i.e. coming from the satellites (satellite clock, trajectory...)

[0009] E_Tropo, respectively E_lono, are the errors related to interferences due to reflections of the SIS on the tropospheric layer, respectively ionospheric layer;

[0010] E Rxsont are the errors specific to the satellite receiver, typically due to measurement noise, phase center and clock bias of the receiver;

[0011] E_Multitrajet are errors due to reflections of the satellite signal on obstacles before reaching the satellite receiver.

[0012] The position error can be modeled and therefore estimated by assuming that all satellites transmit correctly and as a function of the signal-to-noise ratio (SNR); however, this remains a statistical value.

[0013] The types of errors in the satellite-receiver link are also classified as follows, according to where they occur: - "Satellite" errors, due to: clock drift (desynchronization), orbit description (ephemerides), phase center (antenna polarization); - "Propagation" errors, due to: ionospheric delay (layer of the atmosphere from ~80 to 1,000 km altitude, effects related to the presence of free charges), tropospheric delay (layer from 0 to ~80km altitude, effects mainly related to meteorological conditions); - "Reception" errors, due to: measurement noise, phase center, clock bias; - “Local” errors due to: multipathing, decoying.

[0014] There is a need for a reliable location solution that is applicable in particular to the constraints of the railway sector. Summary of the invention

[0015] According to a first aspect, the present invention describes a method for locating a vehicle adapted to move along a set of predefined trajectory(ies),

[0016] a satellite receiver onboard the vehicle being adapted to receive at a instant t of geopositioning signals from a satellite system comprising at least s satellites Sk, k=l to s and s greater than or equal to 3;

[0017] an electronic processing unit comprising a database storing the geographic coordinates Pi of virtual beacons B; distributed along the trajectories of said set of trajectory(s);

[0018] said method comprising the following steps implemented by the electronic processing unit, considering a beacon B; :

[0019] il / for k = 1 to s: calculation of the delay between the received geostationary signal by the satellite receiver at time t from the Sk satellite and the signal from theoretical geostation calculated as expected to be received from the Sk satellite at time t at the position of the virtual beacon Bi as defined by the coordinates geographical Pi of B; in the database, said delay being named correlation delay, X ( S^, t ) ;

[0020] i2 / for the considered beacon B; and the considered time t: calculation of values representative of differences between some of said lags calculated for k = 1 to s and at least one value determined as a function of one or more other of said calculated correlation lags;

[0021]

[0022]

[0023] i3 / estimation of the location of the device based on said representative values ​​of differences between said calculated delays. The proposed innovation uses the correlation principle with best position estimation based on differential measurements of the correlation delay. This approach ensures autonomous measurement integrity, i.e., without the need for additional sensors or installations beyond the GNSS receiver, and is therefore ideally suited for railway navigation. The advantage conferred by the use of correlation delays is that, rather than identifying the nearest beacon as the one maximizing the correlation power (e.g., via ArgnMXB^^Corr(Sh Bj,^ °Û Corr^S^ Bj, t ) “cst 'c maximum of the correlation signal power), we This eliminates the bias due to the fact that the correlation level depends on the received power, while the correlation lag is independent of it. Furthermore, considering differences in correlation lags, and not simply correlation lags, protects against a common bias affecting detection, namely a discrepancy between the receiver clock and the satellite clock relative to universal time. For illustration, [Fig. 3] shows the stability of the differential correlation lag versus the variability of correlation lags in an example where the spacecraft remains stationary. The curve AX (showing on the y-axis the result of the function (MAXi=i to 6 | Max(Xi)(t) -Min(Xi)(t)l)) represents the evolution of the differential correlation lags, and each of Xi, i = 1 to 6, represents the evolution of the correlation lag calculated for a respective satellite Si. The x-axis represents the times of positioning calculation and is expressed in seconds; the ordinate axis represents the correlation delay values ​​(curves XI to X6) and delay gap (curve AX) and is expressed in nanoseconds (ns).

[0024] The invention therefore proposes a reliable positioning solution, making it possible to reduce errors.

[0025] In embodiments, such a method will further comprise at least one of the following features:

[0026] - said representative values ​​of deviations are obtained by at least one of the two The following provisions:

[0027] - calculation of delay differences between Sk, SI satellites:

[0028] Sp Bpt) = Bp t ) -X(SZ, Bp t) |for k , 1 included in 1 to s with k 1;

[0029] - calculation of the standard deviation of the correlation lag as a function of the lags calculated for k = 1 to s;

[0030] - the processing block implements one of the following provisions ila, ilb, ilc, to, considering the Bi tag, implement step 11 of the delay calculation, X(Sk, B, t) for k= 1 to s:

[0031] ila / calculation of the correlation between said received geostation signal and said theoretical geostation signal calculated in B;, said delay X(S^ B{, t) being calculated as a function at least of said calculated correlation;

[0032] ilb / the delay X(S^ Bp t) is calculated, by translation, using the following formula:

[0033] X(St, Bpï) = ^(8^, Bo,t) +j(SB

[0034] where

[0035] SkB0, respectively SkBb is the distance between the satellite Sk and a reference beacon Bo, respectively the distance between the satellite Sk and the beacon b,.

[0036] X^^ B^ t) is the delay between the geostationary signal received by the receiver satellite at time t from satellite Sk and the theoretical geostationary signal calculated as being received from satellite Sk at time t at the position of the virtual beacon Bo as defined by the geographic coordinates PO of Bo in the database,

[0037] c is the speed of light in a vacuum;

[0038] ilc / the delay X(S^ Bp t) is calculated, by translation, using the following formula:

[0039] X(S^ Bpt)=x^ B^i)+ds*co{elk)lQ.3

[0040] where Boest is a reference tag,

[0041] Ala = latitude (Bo) - latitude (B;), Alo = longitude (B0) - longitude (B;) a = arctan (Ala / (Alo * cos (latitude (B0)))

[0042] [3 = azk + a

[0043] ds = d * sin([3]

[0044] where latitude(), longitude() are the coordinates in an orthogonal frame in the local tangent plane;

[0045] - the processing unit implements at least one of the following provisions 4.1, 4.2:

[0046] 4.1 by implementing provision 11a, X(S^ Bm, t) and X(Sk, B02, t) are calculated; then considering BOi as a reference beacon and by implementing one of the formulas in ilb or ilc, the delay in ^02 is determined this time by translation considering Boi as a reference beacon and is named X(SJb t)01;

[0047] and depending on at least the difference between B02, t ) and X(S\, B^, t)Ob the satellite Sk is excluded from the set of satellites taken into consideration for step i3 of estimating the location of the craft as a function of said representative values ​​of differences between said calculated delays;

[0048] 4.2 perform the next step for tags Bi, i = 1 to r:

[0049] by implementing one of the formulas in ilb or ilc: a first delay X(^ B(,Ooi is determined by considering BOi as the reference beacon and a second delay X(S^ B^ t)02 is determined by considering B02 as the beacon of the processing block selects Bi0 with g reference ;

[0050] and depending on at least the differences between X(Sb B;, t)01 and X(S\, B,, t)02 i = 1 to r, the satellite Sk is excluded from the set of satellites taken into consideration for step i3 of estimating the location of the craft as a function of said representative values ​​of differences between said calculated delays;

[0051] - the processing block determines the beacon closest to the device at time t, named Bi0, as by selecting the one of the Bi tags minimizing the dispersion of the values ​​of X( S.. t) , for 1 going from 1 to s;

[0052] - the processing block performs one of the following steps: - the processing block selects with BiQ = ArgnùnB^max ^Sj, Sh Br tp; Var} with - the processing block determines the location of the device as equal to that of a point Rx satisfying a system of equations composed of equations of the following form: X(St Br O-XfS,. B„ t) = Ib^wS's1-LOSs)+ &bSA

[0053] with k, 1 included in 1 to s with k 1;

[0054] c the speed of velocity in a vacuum;

[0055] LO§s resp. LOSst unit vector directed from the beacon Bj towards the satellite Sk , resp. S; ;

[0056] Abs^ = bsk(t) -bs^t), where bsk(O, resp. bsf(O are the errors specific to each satellite Sk, resp. $ / ;

[0057] - the system of equations is solved by choosing pairs of satellites for distinct equations presenting line-of-sight difference vectors ( 1.0^ ; "'lu ^kl

[0058] The local orientation of the track segments on which the vehicle is likely to be located at time t is assumed to be known, and the system of equations is solved by implementing a preliminary step comprising: - selection of pair(s) of satellites (S*, S / ) exhibiting a line-of-sight difference vector orthogonal to said orientation and [LuSs^ ~ LOSsj determination of the value of A ^5 ç as equal to said calculated delay gap B* l)-X(Sb Bb t).

[0059] According to another aspect, the invention describes a computer program intended to be stored in the memory of an electronic device further comprising a microcomputer, said computer program comprising instructions which, when executed on the microcomputer, implement the steps of a process according to the invention.

[0060] According to another aspect, the invention describes an electronic processing unit for the localization of a vehicle adapted to move on a set of predefined trajectory(ies) (VI,V2) and comprising a satellite receiver on board the vehicle to receive at a time t geopositioning signals from a satellite system comprising at least s satellites Sk, k=l to s and s greater than or equal to 3;

[0061] the electronic processing unit comprising a database storing the geographic coordinates Pi of virtual beacons B; distributed along the trajectories of said set of trajectory(s);

[0062] the electronic processing unit being adapted to perform the following operations considering a beacon B; :

[0063] il / for k = 1 to s: calculation of the delay between the geostationary signal received by the satellite receiver at time t from the satellite Sk and the theoretical geostationary signal calculated as having to be received from the satellite Sk at time t at the position of the virtual beacon Bi as defined by the geographical coordinates Pi of B; in the database, said delay being named correlation delay, X(S£, Bb t);

[0064] i2 / for the considered beacon B; and the considered time t: calculation of values representative of differences between some of said lags calculated for k = 1 to s and at least one value determined as a function of one or more other of said calculated correlation lags;

[0065] i3 / estimation of the location of the device based on said values representative of the differences between said calculated delays. Brief description of the drawings

[0066] The invention will be better understood and other features, details and advantages will become clearer from the following description, given by way of non-limiting reason, and from the accompanying figures, given by way of example.

[0067] [Fig-1] Fig.1 schematically represents a railway system putting in works a method of embodiment of the invention;

[0068] [Fig.2] Fig.2 represents steps in a position determination process 100 in an embodiment of the invention;

[0069] [Fig.3] Fig.3 illustrates the stability of the differential correlation gap versus the variability of the correlation lag;

[0070] [Fig.4] Fig.4 represents the values ​​taken by a correlation function in an example;

[0071] [Fig. 5] Fig. 5 represents some of the position determination steps in a particular embodiment of the invention corresponding to Process I;

[0072] [Fig.6] The [Fig.6] illustrates the geometric model considered;

[0073] [Fig.7] Fig.7 also illustrates the geometric model under consideration;

[0074] [Fig.8] Fig.8 represents a local tangent frame of reference used in one of the modes of realization of the invention;

[0075] [Fig.9] Fig.9 also illustrates an example of a marker linked to the tracks;

[0076] [Fig. 10] The [Fig. 10] illustrates the projection of pairs of satellites with a differential line of sight orthogonal to the track segment;

[0077] [Fig. 11] The [Fig. 11] represents the correlation lines.

[0078] Identical references may be used in different figures when they refer to identical or comparable elements. Description of the implementation methods

[0079] With reference to [Fig. 1], a railway system 1 in an embodiment of the invention comprises one or more trains, including train 4 shown in [Fig. 1], which are adapted to move along a predefined trajectory or various predefined trajectories. Each of these trajectories is represented by a series of track segments (for example, left-hand rail segments, right-hand rail segments, or even virtual segments arranged between the left-hand and right-hand rails...).

[0080] Train 4 includes an on-board electronic processing unit 10, which comprises: - an Rx unit 11 including a standard GNSS satellite global positioning system, - an electronic location block 121, - a topographic database 13.

[0081] In [Fig.1], segments of r respective channels Vj, V2, ..., Vr are represented.

[0082] Hereafter, a predetermined position, or reference point, whose spatial coordinates are precisely known, will be called a "virtual beacon" or simply a "beacon", B. The beacons are distributed along the track segments, for example on the line centered between the two track rails, or near the track, for example less than 50 cm from this line.

[0083] For example, the beacons are spaced from their nearest beacons on the same track by a distance equal to or less than D, with D here equal to 4 m for example.

[0084] The track segments and markers are finite in number. Their 3D geographic position (relative to a terrestrial reference frame) and their characteristics are known and are referenced in the topographic database 13. These characteristics include, for example: segment identifier, marker identifier, segment length, segment heading (also called track heading), association between segment identifier and identifier of markers located on (or in the immediate vicinity of) the segment, value of geometric parameters, for example, clothoid or spline, chaining characteristics in the railway network (for example: identifier of successor and predecessor segments).

[0085] The Rx block 11, on board the train 4, is adapted to receive geo-positioning signals from a synchronized local time base, emitted by a set 20 of GNSS geo-positioning satellites: Si 20_l, ..., Ss 20_s, in line of sight of the Rx block 11. Depending on the embodiment, s = 1 or s > Nsat, with Nsat equal to 1, 2, 3, 4 or 5.

[0086] GNSS satellites include, for example, satellites of the American GPS (Global Positioning System) and / or the European GALILEO system and / or the Russian Glonass system, and / or the Chinese Beidou system or any other equivalent system. The positions of the satellites, which may change over time, are predetermined and known to the Rx 11 receiver (via ephemeris data).

[0087] In a known manner, the GNSS receiver in the Rx 11 block is adapted to perform a baseband frequency downsampling from the received signals, to demodulate the signal and extract a geopositioning signal comprising the repetition of a 1023 Hz code. It is this signal in this form which is then referred to as the "received signal" below and which is used, for example, for correlation calculations.

[0088] For example, in a known way, the GNSS receiver is further adapted to extract time and delay information from PRN code (Pseudo Random Noise, used on Galileo and GPS) and to calculate, for each satellite in sight, from this received information, an estimate of the distance between the geolocation device itself and the satellite in sight, also called pseudo-distance.

[0089] The pseudo-distance differs from the actual distance between the satellite in question and the geolocation device due to the errors mentioned above: propagation time estimation errors, due, for example, to atmospheric conditions in the troposphere and ionosphere, and the synchronization error of the geolocation receiver's internal clock. However, it is possible to eliminate common errors (including the receiver's time bias) by relying on information transmitted by a plurality of separate satellites.

[0090] For example, the PRN code of the C / A (Coarse Acquisition) signal in the case of GPS is a digital signal composed of 1023 chips and repeats every millisecond. It should be noted that the term "chip" used in GNSS techniques is distinct from the term "bit," which is used to define a unit of information. The code has a given code length, denoted L(CI A). Its time length is 1 ms, and therefore its spatial length is 1 ms * c, where c is the speed of light. From this, we deduce a chip time length of 1 ms / 1023.

[0091] The localization block 121 is adapted to estimate the position of the train 4 considered at time t.

[0092] Depending on the embodiment, this location estimation takes different forms. For example, in one embodiment, it includes determining the position of the estimated beacon closest to train 4 and / or the geographical coordinates of the Rx receiver; in another embodiment, it allows the validation or not of a previously determined positioning hypothesis; in another embodiment, it allows the identification of the track segment on which the train is a priori positioned (being certain of the train's track is indeed a major safety element to avoid accidents).

[0093] The location estimation is carried out, according to the invention, based in particular on the satellite positioning signals received by the Rx 11 block in its position to be estimated, and furthermore, in embodiments, based for example on measurements from an on-board inertial measurement unit and / or based on measurements by an odometer.

[0094] Method for determining position according to the invention, based on correlations between predicted signals relating to virtual beacons and received signals

[0095] With reference to [Fig.2], a method 100 for determining the position of the train 4 at time t is implemented by the processing block 10 in an embodiment of the invention; it comprises a set of steps 101 to 105, exploiting the received GNSS signal and the prior knowledge of the positions of the fixed and predetermined virtual beacons B provided by the topographic database 13.

[0096] Such a process exploits the fact that the trajectory of train 4 necessarily follows the segments of railway track, which are known and on which there are markers.

[0097] The processing is implemented by correlation of the GNSS signal received, by the Rx block 11, at the local reference date t of the GNSS receiver, from the various visible satellites Sk, k = 1 to s (s being according to the embodiments the maximum number of available satellites or a subset), with each of the expected code signals (replica code signals) from these satellites at the same local reference date t of the GNSS receiver, and calculated for the position P; of a virtual beacon B;.

[0098] Depending on the case, the beacons considered B;, i = 1 to N, are either the set of beacons B, or a strict subset of this set of beacons B (therefore comprising a reduced number of beacons and determined taking into account certain basic assumptions considered known: train 4 located on a stopping track or a departure track or a running track, in a restricted geographical area, etc.). The beacon B; is located at P;, a 3D geographical position specified in association with B; in the base 13.

[0099] In one embodiment, the processing block 10 comprises a processor and a memory, in which software instructions are stored, which, when executed on the processor, implement the steps of the process 100.

[0100] Thus, in step 101, the theoretically received signals (replica signals) at time t at the level of the beacon B;, (t), from each Sk satellite in view of the beacon are predicted by the location block 121 (because as the position of the beacon is known, the travel time of the satellite signal to the beacon is known).

[0101] During the generation of local GNSS replicas of the expected codes, errors are taken into account using available error models, i.e., the known clock error, ionospheric error, and tropospheric error models. These error models can be provided by the onboard GNSS receiver or via an assist link, and allow errors in the expected satellite distances to be reduced to a few meters.

[0102] It is assumed that the residual errors on the code delay after application of the corrections are centered Gaussian (in one embodiment, the validity of this assumption is then tested in an integrity proof).

[0103] In step 102, at time t, the RX11 block receives, in its current position, the signals (t) sent by the Sk satellites, k = 1 to s and provides them to the location block 121.

[0104] In step 103, for beacon B; and each satellite Sk, the correlations between predicted signals Ccafc ( t ) and signals Cg^sk ( C received at time t by the Rx block 11 are calculated by location block 121.

[0105] The correlation functions are calculated for the B beacon; and for each Sk satellite - at least for a certain number of values ​​taken by ':

[0106] rg T ) = < t _ T )

[0107] (the integration interval is for example fixed at the value T, typically equal to at least the duration of a navigation bit, i.e. 20 ms or 20 PRN code lengths).

[0108] In step 104, for beacon B; and for satellite Sk, the location block 121 then estimates the value of r in which the correlation function D^(r ) reaches its maximum; this value, which is the correlation lag between the two Ceale BiSk ( O and Cgxsk (t) signals, is named X( Biy t ).

[0109] The code delay X(S^ B* t) is actually composed of a propagation error term and residual error term, as well as a geometric term dependent on the distance between the beacon and the receiver, and on the azimuth and elevation of the satellite considered.

[0110] In step 105, for the considered beacon B; and for the considered time t: the location block 121 calculates representative values ​​of deviations between some of said calculated correlation delays and at least one value determined as a function of one or more other(s) of said calculated correlation delays, and the location block 121, as a function of these deviation values, estimates the location of train 4.

[0111] The use of this differential method eliminates the receiver bias during the beacon detection phase. This results in greater reliability in beacon detection. Optionally, processing block 10 estimates this bias after beacon detection, before potentially applying an integrity algorithm.

[0112] Example of position search when the train is stopped, on a siding or departure track (static case):

[0113] In such a case, for the implementation of algorithm 100, only the subset of base 13 markers identified therein as being located on segments of sidings or departure tracks are considered. These subsets of markers are therefore, in this case, the markers B;, i=l to N.

[0114] Example of position search when the train is moving on a track (dynamic case):

[0115] In such a case, for the implementation of algorithm 100, only the subset of base 13 beacons identified therein as being located on traffic tracks is considered (this subset can be further reduced by retaining only the beacons located on tracks on which the train is likely to be located, this (assessment being determined, for example, based on a maximum displacement evaluated according to measurements from an on-board inertial measurement unit and a previous reliable position of the train). These subsets of beacons are therefore, in this case, the beacons B;, i=l to N, considered.

[0116] The dynamic case differs from the static case in that, in addition, the correlation is calculated at different times, with the correlation being maximum when the train is at the beacon. Indeed, in addition to beacon detection (maximum spatial correlation), it is also necessary to detect the time at which this correlation is maximum (which is not necessary in the static case).

[0117] Fig. 4 presents a one-dimensional graphical representation, as a function of time represented on the abscissa, of a correlation function mentioned above, schematically represented in the form of a triangle.

[0118] The duration 24 is typically the duration of a chip of a GNSS code sequence.

[0119] This correlation function has a maximum of 26, as determined, The abscissa tijk is therefore the code delay. In some embodiments, an interpolation is also performed to better estimate the position of the maximum.

[0120] In embodiments, when calculating the values ​​of the correlation function for the beacon B; and the satellite Sk, i = 1 to N and k = 1 to s, the following steps are carried out (using in particular three multiplier modules and three integrator modules): - a convolution product is performed: the received satellite signal C^xk(O is multiplied with the replica signal (t) which is the a priori perfectly aligned replica of the code (if the train is positioned in Pi and in the absence of uncompensated errors), then the result is integrated: the output value, named Prompt (P), is representative of the correlation between these two signals; - the received satellite signal CrxJc ( ) cst multiplied with the advanced replica signal Ccalc Bik (£ + <?), puis 'c résultat est intégré : la valeur de sortie, nommée Early (E) est représentative de la corrélation entre ces deux signaux ; par exemple la valeur 5, positive, par exemple est égale à 7,5 ns (l’idée étant d’avoir un intervalle Early-Late prédéfini qui est représentatif du besoin en précision, notamment la détection de voie ; les voies étant considérées espacées de 4m, soit 15ns en temporel, on choisit par exemple ici un écart de 15ns entre les points Early et Late). - The received satellite signal CrxJ< (O constant) is multiplied by the delayed replica signal Ccak Bijc ( _ , then the result is integrated: the output value, named Late (L) is representative of the correlation between these two signals; - in embodiments, in order to obtain more points of the correlation function, additional advanced and delayed replica signals of the shift values ​​51, 52, are for example also multiplied and integrated. In what follows, we assume that Doppler is perfectly understood. Procedure I

[0121] The closest beacon to train 4 is a priori the one that minimizes the average of the squares of the X^ Bb t) i.e. .4^^^,(^=1^(¾^^) 2)-

[0122] The applicant further determined that the beacons closest to the Rx 11 receiver were those which minimized the dispersion of the values ​​of X(S}, B,, t) between satellites.

[0123] Thus the virtual beacon estimated as that corresponding to the location of train 4 being that which minimizes the maximum differences in relative code delay between the satellites considered, in an embodiment of process 100 considered here corresponding to Process I, the localization block 121 is adapted to implement process 100 by carrying out during the implementation of step 105, the steps described below with reference to [Fig.5].

[0124] In a step 105_la, for each of the beacons B;, i = 1 to N, the location block 121 calculates the delay differences between satellites Sj, Sk, for j, k included in 1 to s with kj:

[0125] =\X(Sj, B, tyX(Sh B, t) |

[0126] In a step 105_lb, the localization block 121 determines the nearest beacon to train 4 at time t, named Bi0, by selecting the one among the beacons B;, i = 1 to N, minimizing the dispersion of the values ​​of X(S;, B^ t) between satellites, i.e.:

[0127] B^ = Argmins^maxô^Sj, Bb t

[0128] The minimum number s of satellites considered for this method I is 5, in one embodiment.

[0129] In one embodiment, the track segment of which the beacon is a part according to base 13 is further determined by the processing block 10 as the track segment on which the train 4 is positioned at time t.

[0130] In another embodiment, the localization block 121 will use, instead of, or in addition to, these delay differences ô(Sj, Sk, B;, t) between the satellites for a beacon B;, the standard deviation (or variance) of these delays, which similarly represents the dispersion of lags. For example, for each considered beacon B;, the variance Var^i^S*, B, t)= B,. t) is calculated, the The closest determined beacon by block 121 is then the one, among the beacons Bi to Bn, whose variance is minimal. Geometric model

[0131] The correlation delay X ( Bj, t ) corresponds to the code delay measured between the signal received by the receiver Rx from the satellite Sk and the signal predicted at the level of a beacon Bj considered.

[0132] In the absence of propagation error, satellite error, receiver error and local errors, Bp t) is therefore the difference in travel time between the satellite-beacon path and the satellite-receiver path.

[0133] We then have xej for the Sk satellite and the Bj beacon. With X ( Sk > Bj 1 ) = Bj ~ SkRx) F From a geometric perspective, this relationship can be expressed to the first order as a function of the distance between the receiver and the beacon.

[0134] With reference to [Fig.6], by naming: - Hk the orthogonal projection of the position of Sk in the local plane tangent to the earth at the location considered (and for example associated with the NED (“North, East, Down”) or ENU (“East, North, Up”) coordinate system), - elk is the elevation of the Sk satellite and azk' is the relative azimuth of the Sk satellite with respect to the line between the Bj beacon and the Rx receiver, we have:

[0135]

[0136] y / ç » X(Sk,Bj, t) - ASjBj-Si.Rx] -

[0137] c being the speed of light in a vacuum.

[0138] Applying the law of cosines in triangle HkBkRk, we obtain:

[0139] HkRx = ^HkBj + RxB2j - 2.cos (azk) JIkBjRxBj

[0140] I HkRx = HkB^ 1 + - 2cos (azk)

[0141] The radical expands to the first order, which gives: HkRx « HkBj - cos (az'k)RxBj

[0142] and now considering the equation above of X^Sj? Bj, t):

[0143]

[0144] This approximation can be physically considered an equality given the difference in order of magnitude between the distances SkBj and RxBj. It follows that the delay of the received signal relative to the predicted signal can then be written as:

[0145] :

[0146] X^Bj, / ) = co^az'^RxB ; C£os(el^

[0147]

[0148] Case of relative azimuth: The relative azimuth, ,7~7r .T"?,*- can be expressed as a function of the azimuth azk of the azk = BjRx,BjHk satellite Sk and runway heading 0 if receiver Rxest located on the same track as beacon Bj: az'k=6-azk (see [Fig.7]).

[0149]

[0150] t) = cos^O-az^JtxB j c.cos{el^)

[0151]

[0152] This expressed form can be written using a dot product as follows: . „ s coffaz)HtBi i---a——.

[0153] where LOSsk is the line of sight vector (“Line Of Sight”), of unit norm, starting from the beacon Bj (or the receiver Rx, given the distances) and pointing towards the satellite Sk. Measurement model

[0154] The measurement of X( S^, Bj, t ) (i.e., the result of the calculation performed in step 104 as a function of the received signal) is affected by errors related to signal prediction and the correlator, coming from several sources including:

[0155] - error in correcting / modeling the propagation of the received signal (ionosphere, troposphere);

[0156] - satellite errors (orbits and satellite clock) impacting weather prediction of emission and satellite position;

[0157] - receiver errors (clock, group delay) for time estimation emission;

[0158] - local errors (multipath, spoofing) lengthening the signal propagation time received ;

[0159] - correlator sensitivity / resolution and correlation noise.

[0160] These errors have spatial and temporal properties that differ, and influence the value of X(Sk, B j, t) differently.

[0161] Ionospheric and tropospheric correction errors are added directly to the geometric travel time difference, and are locally the same for all beacons.

[0162] By grouping together, on the one hand, receiver errors common to all satellites (BREC) and, on the other hand, errors specific to each satellite (orbit, satellite clock, ionospheric and tropospheric model error, multipath, ...) that we Let bsk be the value denoted by X^S^ Bj, t ) the measured value (i.e. calculated from measurements of the received signal and therefore incorporating errors):

[0163] X(ShBjJ>) ~^BjRx.L03sk+brec + bsk

[0164] If we differentiate between 2 distinct satellites Sk, Si, i.e., we calculate X ( Sb Bj,t ) - X ( S[, Bj, t ) , we obtain:

[0165] Ô^S^Bj,^ ^X^Bj,^ - XtS^t) = ^B]Rx(LÔS^k-LOSs) + bSk(t)-bs,(t)

[0166] The unknowns in this equation are the coordinates of the receiver (point Rx) (i.e., what we want to determine) and the biases bsk(t), b^(t). By considering the bias differential A bs^(t) = bsk(t) - bs^(t) instead of the biases, this reduces the number of unknowns. Procedure II

[0167] In an embodiment of the process 100 now considered corresponding to Process II, as indicated in the statement of the general principle of the invention, the measurements of X(Bj, t) obtained in step 104 are differentiated to cancel the errors due to the receiver brec (mainly due to desynchronization effects) and these differences are used to estimate a position based on the model described above, the localization block 121 being adapted to, during step 105, implement the resolution of a system of equations, named SYS, to determine the coordinates of the receiver, the system comprising a sufficient number of equations of the type: 101681 5~x( S* S, Bj, t ) = ^B^LÔS^t - LOSs) + A bSA

[0169] the ôx(Slp S / , Bj, t) being taken equal to the result of the difference between ^(¾ B(, t) and X(S[, Bj, t) from step 104 (therefore in reality "measured"), considering several distinct pairs of satellites with k, 1 included in 1, .. ,,s and k 1 and the same beacon Bj, hereinafter referred to as the reference beacon.

[0170] The minimum number s of satellites considered for this method I depends on the assumptions taken for its resolution: s equal to 3 allows obtaining 3 pairs of satellites and solving the system to determine the coordinates of Rx in a three-dimensional frame.

[0171] Any method of resolution can be used to solve this system, each equation having as unknowns the coordinates of Rx and the difference in bias.

[0172] Advantageous methods for solving the system of equations are further described below, either alternative or cumulative, depending on the embodiment. In some embodiments, the locating block 121 is adapted to implement one or both of these methods.

[0173] First, the dot product can be calculated in 2 dimensions instead of 3, by expressing the vectors in the local tangent frame to which the receiver belongs (instead, for example, of the classically used ECEF (“Earth Centered Earth Fixed”) frame): such a local tangent frame (O, E, N, Z) of the NED or ENU type is shown in [Fig. 8], where any point P in space is represented by its coordinates (e, n, z). Furthermore, all points on the map, and the position of the receiver's antenna after applying the lever arms, are considered to be located in the same plane.

[0174] By choosing the reference beacon Bj as the origin of this reference frame, we can write: [0!75] Sb Bj, t ) = z(nRn^osSA + eReALOsSlis) + A bs^

[0176] where e., respectively n., are the coordinates on the E axes, resp. N, of the element indicated as a subscript (thus eR* is the coordinate on the E axis of Rx)

[0177] With reference to [Fig. 9], by further choosing a coordinate system linked to the track (X, Y), we can move, for the coordinates of receiver Rx, from 2 continuous unknowns to 1 discrete unknown and 1 continuous unknown (this coordinate system allows us to take full advantage of the fact that the train is located on one of the tracks). We assume that all the tracks are locally parallel. X is the continuous longitudinal axis and Y the discrete transverse axis. The above equality can then be written as: + Abs^

[0179] Pas is the distance separating two tracks, nbvoie is the discrete value of the track on which the train (i.e., the receiver Rx) is located, XMÆSs,^ resp. y&LOSss,, is the coordinate of the vector r „J””), on the X axis, resp.Y and X^x is the coordinate of the position of the [LOSsk - BOSsJ receiver Rx on the X axis.

[0180] The lane detection problem will consist of finding the discrete value of the Y-axis coordinate. Note that by making a lane assumption we obtain a pair (longitudinal coordinate; bias differential).

[0181] One solution method, considering the reference beacon Bj, is to make different track assumptions, solve the system for each assumption using several satellite pairs, and choose the assumption with the minimum bias variance. This solution is particularly useful in railway applications because it reduces the error space.

[0182] One method for solving the system of equations is to use several receivers and to use the known distance between these receivers (static cold start case in a straight line). S(, Bj,t)=

[0183]

[0184]

[0185]

[0186]

[0187]

[0188]

[0189]

[0190]

[0191]

[0192]

[0193] One method for solving the system of equations involves considering differential biases nu Abs for certain pairs of satellites. To solve the In a system without a track assumption, this condition is imposed on two pairs in one embodiment. Reconsidering the embodiment with passage through the track frame, we then consider the following system of equations with the satellite pairs 5?) [ MS S2, Bj, / ) = Ünbvoi*pas*^^ +xrx^loss s \ J \ 1 7 S3, Bj, t) Ilnbva. *pas*yAUJS + xRxALossA \ .>'1 * 7 The system can be solved with 2 pairs. It may be interesting to observe the differences in detection depending on the pairs chosen. The matrix representation of the unbiased system is: M = HRX with r) ^(S2,S3,Bj,t). the measurement vector XALOSSlS, y^LOSg^ and -¾ XaLOSs s, yALQSs s X yRx ~P^ Rx is therefore obtained by multiplying M by the inverse of H. The inversion is for example carried out by optimization constraint, or by successively taking several path hypotheses, and comparing the longitudinal x values ​​thus obtained. This resolution is, for example, a useful preliminary step for a subsequent step where certain differential biases are estimated, and a choice of pairs for which the bias is not estimated will have to be made. The result can also be compared in an embodiment to the beacon detected as the closest (for example with Method I or another solution), to validate / invalidate this detection or to help in the choice of pairs requiring estimation. Geometric conditions for choosing the pairs: the ability to solve this system will also depend on the observation matrix M. We can assume that choosing two pairs of satellites with orthogonal line-of-sight difference vectors ( ) and TT „2^ maximizes observability (and this LOSs-LOSs) LOSs^-LOSsJ independently of the method chosen to solve the system of equations SYS). In one embodiment, only pairs of satellites $ ) and (S7 S3) For which the absolute value of the determinant of H is greater than a threshold h (for example h = 0.5) are retained to establish the equations of the system.

[0194] In one embodiment, for track detection, the asymmetry of the problem and the fact that the track direction is perfectly known can be exploited (i.e., it is assumed here that all track segments on which the train is likely to be located have substantially the same direction). In particular, it is observed that satellites whose line of sight is perpendicular to the track line provide correlation delay measurements X() without geometric error, the dot product then being equal to zero. The same logic is applied to B^LOSs^LOSs) pairs of satellites $ ) whose differential line-of-sight vector, i.e / , is orthogonal to the track line: Figure 10 illustrates this situation [LOSsk - LOSsJ for couple n°1 <^) and couple n°2 ^).

[0195] For a pair S{ of satellites satisfying these conditions, the dot product is equal to zero for the beacons Bj located on the channel, it is then obtained: 101.61 4(S„ S„ Bj, 0 = JB^LoSs,-LOSs)+b*(t)-b^W =bs,(t) -bSl(i)

[0197] For example, this measurement is then averaged with the 8% calculated for other beacons on the same channel in order to reduce the noise effect of the model. Averaging with other beacons does not change the differential bias but reduces the measurement error due to noise. These arrangements therefore make it possible to estimate the differential bias corresponding to the satellites and to reintroduce this value into the SYS system of equations to deduce the receiver's position.

[0198] For the other channels (indeed, the channels to which the receiver does not belong introduce a resulting geometric bias) a geometric component is added, theoretically equal to mn where A is the distance between £11 A 11 two lanes (assumed constant) and m an integer to cover cases of distant lanes (not just adjacent to the true lane). The value of in / ""xu can ê A. | J \ LOSsk - LOS s J 11 can also be estimated by double differentiation using beacons from two different channels, using the same pair of satellites (Sk; SJ.

[0199] In one embodiment, to detect the track on which train 4 is located, a simple method consists of selecting the track, among the considered tracks, corresponding to the minimum average value ôA(L l, Bj, t) (the average being calculated over the Bj values ​​of the same track). In practice, this often gives a good result, mainly because the residual biases are positive and provide a weaker bias being differential (however, a biased couple can trigger a poor lane detection (the worst case scenario would be that the differential bias perfectly compensates for the geometric resultant).

[0200] This method provides a lower bound for the residual bias, equal to the minimum mean ôx(k, L Bj, t). Furthermore, to have a channel error, the bias must compensate for the resulting component: ,, , xt To n / When choosing a satellite with a O) >Mr. W-WSs; iatj » i \ ai / il With a higher differential LOS standard, the probability of having a differential bias exceeding this threshold is reduced and can provide more confidence in the detection. Therefore, in an embodiment of the perpendicular case, at least one, some, or each of the pairs of satellites considered will be chosen, exhibiting the highest values ​​of h. It is also noted that the level of confidence to be given to a LOS (k-LOSS1) The torque perpendicular to the path varies according to the azimuth and elevation of each satellite.

[0201] To increase the confidence of the detection, another pair with a differential LS orthogonal to the path is used in one embodiment. If the differential bias is assumed to be positive (the ability to detect that one satellite is more biased than another), taking a pair with a differential LS of opposite sign to the first allows the result to be crossed, and if the bias is assumed to be positive, with the bias having an effect on the differential LS, one should expect to have a common intersection for the true path, and unjoined paths for increasing biases, "one pushing and the other pulling".

[0202] With two pairs, we then find that we have higher protection, the case of poor detection requiring a differential bias greater than the resultant for both pairs and an error in the assumption of positivity for one of the two pairs.

[0203] Furthermore, it is noted that if the correlation lag, between tags this time, is differentiated, it is obtained:

[0204] XS^Bj,t)-X(Sk,Bbt) = ?BfixLOSSk-tB^LO^ = tB^LO^, which is a term independent of the position of the Rx receptor. This formula can be used by processing block 10 to translate the result of Xmax to all the beacons, and to verify the consistency of the measured correlation delay values, which reduces the number of calculations performed and / or increases the reliability of the results by limiting the beacons considered only to those corresponding to a value XB^-XtS^Bt, t) consistent with all other values ​​of this type.

[0205] Residue-based protection method

[0206] A technique takes into account the variance of the residual biases for each position hypothesis. Each residual bias vector for a position hypothesis can be linked to another residual bias vector by a translation vector.

[0207] Indeed, equality

[0208] BjJ) = ±B]^LÔ^^-bSlW

[0209] can also be written:

[0210] with Rvrai the actual position of the receiver and Besümé the estimated position of the receiver.

[0211] The residual bias estimation error can be considered to be related to the position estimation error by c Bes(jm^Rvraf.(LOSsk "

[0212] Note that if we place ourselves in a frame of reference linked to the track (longitudinal axis = track axis), if the vector LO&, _LO§s is perpendicular to the track axis, the residual bias error depends only on the track error. If this vector is parallel to the track, the error depends only on the longitudinal position error. And between two position hypotheses, we have a translation relation on the residual bias error whose translation term is known.

[0213] In the preceding cases, the residual bias was estimated via a position hypothesis by [02141 bs,w-bs^t) =6^1. Bj,t)4Bj R^JlOSSi-LOSs)

[0215] By applying a threshold to the bias vectors, it is possible to reject positional assumptions by rejecting those assumptions for which the norm of the bias vector exceeds the chosen threshold. This threshold can be calculated based on Gaussian distributions for the residual errors. This protection method is implemented, for example, by an electronic control unit 122 within the processing unit 10.

[0216] Alternative solutions for determining the values ​​of delayX] t )

[0217] There are different solutions for determining the delay values ​​B^t).

[0218] One of these solutions was described in steps 101 to 104 considering the beacon Bi and the set of satellites Sk, k=1 to s, according to which the correlation of the signals received and theoretically received at B is calculated, and the delay B,, t) is deduced from this calculated correlation. Typically, if we consider P possible positions (i.e., P beacons), S satellites, and 5 time offsets for the correlation, then we will have P*S*5 correlation points to be calculated (with for each correlation point, a calculation of the predicted code and the correlation with the signal actually received, then for each position and each satellite, a delay Xmax is calculated based on the result of correlation of 5 time shifts using a parabolic interpolation: thus a large volume of calculation is implemented.

[0219] In one embodiment, the invention implements, instead, a solution for determining the delay values ​​X(Sk. t) according to one of the two approaches described below, exploiting the relationship between the delays relative to one beacon and the delays relative to another beacon, and thus allowing the correlation to be calculated only for a reference beacon.

[0220] Two approaches are proposed here.

[0221] Approach 1:

[0222] The correlation calculated with respect to one of the beacons, for example the BH beacon for the Sk satellite, can be expressed as follows:

[0223] ^Tint J ro CRXSk( O • Cculc ELSk^ )

[0224] It follows that between two beacons, Bb B2, the correlation difference is simply the difference in the predicted signals Ccuic B(t), C(.a}cB2Sk(t), all the beacons using the same error correction for predicted signals from Sk, it It follows that:

[0225] CcalcBLSk(O - CcaleB2JSk(^

[0226] c is the speed of light in a vacuum; where SkBb respectively, is the distance between the satellite Sk and the beacon Bb respectively the distance between the satellite Sk and the beacon B2.

[0227] Using this approach, the location block 121 determines the delay t) of any beacon B; by "translation", instead of steps 101 to 104, from the delay determined for a reference beacon Bo:

[0228] t) = 0 + 2^ g. _ SkBoj'

[0229] where SkB0, respectively SkB,, is the distance between the satellite Sk and the beacon Bo, respectively the distance between the satellite Sk and the beacon Bj.

[0230] Approach 2:

[0231] This approach 2 is based on the concept of a correlation line. According to the correlation formula, at positions (i.e., between beacons) that are equidistant from the Sk satellite, there is the same predicted signal and therefore the same correlation result. These positions (beacons) lie on the same line on the ground, called the correlation line. The difference between the relative delays of any two beacons is, in fact, the distance between these correlation lines on which these two beacons are located. Figure 11 illustrates, among the set of points indicating beacon positions, the correlation line on which beacon Bo is located and the correlation line on which another beacon, B, is located; and the distance, ds, between these two correlation lines.

[0232] Using this approach, instead of steps 101 to 104, the location block 121 determines the delay B» t) of any beacon B; by "translation" from the delay determined for a reference beacon Bo using the following formula

[0233] Bi,V)=x(Skl B^i)+ds^4elk)lQ3

[0234] where

[0235] azk, elk: azimuth angles, respectively elevation angles of the Sk satellite;

[0236] a: angle between the line passing through Bo and B; and the axis E;

[0237] [3 : angle between the line passing through Bo and B; and the correlation line passing through B; ;

[0238] d: distance between the Bo, B beacons;;

[0239] ds: distance along the line of the Sk satellite in the plane, between the Bo, B beacons;( In other words, it is the projection of d according to SkB0)

[0240] the following calculations being performed:

[0241] Ala = latitude(Bo) - latitude(B;), Alo = longitude(B0)- longitude(B;), the coordinates called longitude(), latitude() of the Bo, B; be given in a NED frame (or in any orthogonal frame also in the local tangent plane);

[0242] a = arctan(Ala / (Alo *cos(latitude(B0)

[0243] [3 = azk + a

[0244] ds = d * sin([3]

[0245] and finally: X(S^ B, t) = x(Sh t) +rfs*cos(^) / 0.3

[0246] Implementing either of these two approaches makes it possible to significantly reduce the computational load required to implement the localization solution according to the invention.

[0247] Tests were carried out using real signals collected from the channels, allowing the correlation delay X() to be deduced for seven beacons by "translation" from another reference beacon. The difference between the delays obtained by translation and those actually measured by correlation calculation was calculated: it is approximately 0.1 m on average and a maximum of 0.39 m, and the two approaches give comparable results with an average difference of 0.02 m and a maximum difference of 0.1 m.

[0248] Furthermore, it is necessary to use precise values ​​for the correlation delays in the localization solution according to the invention. However, the precision of the correlation delay X() obtained based on correlation calculations using The measurements of received signals and calculation of theoretical signals at the beacons. Inaccuracy can come from noise or interference on the GNSS signal and / or interpolation errors.

[0249] The translated values ​​of correlation delay are used to evaluate the accuracy of the "measured" correlation delays and to validate / invalidate satellites: see the two examples of embodiments described below.

[0250] For example, considering two beacons BOi, B02 (13.5 m apart), the X(SÀ„ Bp l) “measured”, i = 01, 02, is determined by effective correlation calculation (from measurements at these beacons) relative to these beacons, by implementing steps 101 to 104.

[0251] In a first validation method, the comparison of correlation delays at one or more beacons obtained by translation (using one of the two approaches described above) and those actually "measured" at this / these beacon(s) (using steps 101 to 104) is used to evaluate the accuracy of the "measured" correlation delays and to validate / invalidate satellites. Starting from this X(S^ B;, t) thus calculated (measured), i = 01, by translation this time from the reference beacon BOi, a translated Bp t)i=02 is determined. The differences between the measured and translated X(S^, Bj, t )02 are then compared, and if the difference exceeds a predetermined threshold, Sk is excluded from the set of satellites used by the localization block 121 to implement the localization determination in step 105.

[0252] In a second validation mode, each of the BOi, B02 tags is used as a reference tag by the locating block 121 to calculate by translation, by one of the two approaches described above, the correlation delay X(Sfc, Bj, t), j = 1 to r.

[0253] So for each beacon Bj, j = 1 to r, the location block 121 obtains a first value X(S^, Bj, t) Oi by translation from the measured ^(¾ Bm, t) and a second value X ( Bj, t ) 02 by translation from the measured X ( S / r B02, t ), The differences between X ( Bj, t ) oi^tX ( Slc Bj, t ) 02 translated are then compared, j = 1 to r and if the difference exceeds a predetermined threshold (specific to the satellite Sk or common to different satellites), Sk is excluded from the set of satellites used by the location block 121 to implement the location determination in step 105.

[0254] It should also be noted that measurements of the codes in the positioning signals have been described above. The invention can also be implemented by measuring, instead of—or in addition to—the phases of the carriers of the positioning signals.

[0255]

[0256]

[0257]

[0258]

[0259]

[0260] The method according to the invention is adapted to the static case, for consolidating channel detection at initialization and during a "cold start," and to the dynamic case, provided that the integration times for correlation calculations are adjusted. The time of passage over the beacon in the dynamic case could be an additional parameter to refine position detection. A robust and simple version can do without this parameter by using regular and fixed detection times, thus detecting the nearest beacon for a similarly estimated position detection time. In other embodiments, the invention is implemented using a different formula for X(t) in the geometric model, for example using: XÏS^ Bj, t) = = —1r-'B^RxIOS^t \ & J' J c£ot{elk) c^cosWj J Method 100 can be implemented by executing software instructions on a processor, as described above. Alternatively, either method can be implemented by dedicated hardware, typically a digital integrated circuit, either specific (ASIC) or based on programmable logic (e.g., FPGA / Field Programmable Gate Array). The method according to the invention is of course usable, beyond the railway field, in any system in which it is necessary to estimate the position of a vehicle whose movement is only possible on a predetermined and finite set of trajectories, a database storing the geo-referencing of virtual beacons located on the trajectories.

Claims

Demands

1. A method for locating a craft (4) adapted to move along a set of predefined trajectory(ies) (VI,V2), a satellite receiver onboard the craft being adapted to receive at time t geopositioning signals from a satellite system comprising at least s satellites Sk, k=l to s and s greater than or equal to 3; an electronic processing unit (10) comprising a database (13) storing the geographic coordinates Pi of virtual beacons B; distributed along the trajectories of said set of trajectory(ies); said method comprising the following steps implemented by the electronic processing unit (10), considering a beacon B;i1 / for k = 1 to s: calculation of the delay between the geostationary signal received by the satellite receiver at time t from satellite Sk and the theoretical geostationary signal calculated as expected to be received from satellite Sk at time t at the position of the virtual beacon Bi as defined by the geographic coordinates Pi of B; in the database, said delay being called the correlation delay, X(Sfc t); i2 / for the beacon considered B; and the time t considered: calculation of representative values ​​of differences between some of said delays calculated for k = 1 to s and at least one value determined as a function of one or more other(s) of said calculated correlation delays; i3 / estimation of the location of the device as a function of said representative values ​​of differences between said calculated delays.

2. Localization method according to claim 1, wherein said representative values ​​of deviations are obtained by at least one of the following two arrangements: - calculation of delay deviations between satellites Sk, SI : = 1^(¾ 0 1 for k, 1 included in 1 to s with k 1 ; - calculation of the standard deviation of the correlation delay as a function of the delays calculated for k = 1 to s.

3. A localization method according to claim 1 or 2, wherein the processing block implements one of the following arrangements: ilb, ilc, for, considering the beacon Bi, to implement step il of calculating delays, X( t ) for k= 1 to s: ila / calculation of the correlation between said received geostation signal and said theoretical geostation signal calculated in B;, said delay Bb i) being calculated as a function at least of said calculated correlation;The delay X(SJt, B,, t) is calculated, by translation, using the following formula: X(S^ B^ l ) = X(Sfc B^ t) + g _ where S^B^ respect ivcmcntSkBH is the distance between the satellite % and a reference beacon Bo, respectively the distance between the satellite Sk and the beacon B, XGSA, 0 is the delay between the geostationary signal received by the satellite receiver at time t from the satellite Sk and the theoretical geostationary signal calculated as having to be received from the satellite Sk at time t at the position of the virtual beacon Bo as defined by the geographic coordinates PO of Bq in the database, c is the speed of light in a vacuum; The delay X(Si, Bh t) is calculated, by translation, using the following formula: B^t) = x(Sh where Bo is a reference beacon, Ala = latitude(Bo) - latitude(B;), Alo = longitude(B0)- longitude(B;)a = arctan(Ala / (Alo *cos(latitude(B0) )) [1 — azk + and ds = d * sin([3) where latitude(), longitude() are the coordinates in an orthogonal frame in the local tangent plane.;

4. Localization method according to claim 3, wherein the processing block implements at least one of the following provisions 4.1, 4.2: 4.1 by implementing provision ila, X(S^Bm, t) and X(S^BO2, t) are calculated; then, considering BOi as the reference beacon and by implementing one of the formulas in ilb or ilc, the delay in Æ02 is determined this time by translation considering Boi as reference beacon and is named Bü2, t )01 ; and as a function of at least the difference between X(S^, B02, t) and X(S*, B(a, l)Ob the satellite Sk is excluded from the set of satellites taken into consideration for step i3 of estimating the location of the craft as a function of said representative values ​​of differences between said calculated delays; 4.

2. Carry out the next step for beacons Bi, i = 1 to r: by implementing one of the formulas in ilb or ilc: a first delay B^t)01 is determined by considering BOi as a reference beacon and a second delay X ( B^t ) 02 is determined by considering B02 as a reference beacon; and as a function of at least the differences between X(S&, B^t)01 and Bp t)02 i = 1 to r, the satellite Sk is excluded from the set of satellites taken into consideration for step i3 of estimating the location of the craft as a function of said representative values ​​of differences between said calculated delays.

5. A localization method according to any one of the preceding claims, wherein the processing block determines the nearest beacon to the craft at time t, named Bi0, as by selecting that of the beacons Bi minimizing the dispersion of the values ​​of X( Bj, t) , for 1 ranging from 1 to s.

6. A localization method according to claim 5, wherein the processing block performs one of the following steps: - the processing block selects Bi0 with Bi0 = ArgminB^rnax jjk ^Sj, Bb t; - the processing block selects Bi0 with Bn = ArgminBiyar) with Vari = UX^' = B„ t)

7. A localization method according to any one of claims 1 to 4, wherein the processing block determines the location of the device as equal to that of a point Rx satisfying a system of equations composed of equations of the following form: X(Sh B, t)-X(Sz, t) ^lB]Rx.(LOSsk-LOSs)+ AbSkSl with k, 1 included in 1 to s with k ≥ 1; c the speed of the velocity in a vacuum; LO$S, resp. lOSs unit vector directed from the beacon B in the direction of the satellite Sk, resp. S; A — bs, (t) - bs,(t), where bg. (t), resp. bs,(Z) are the errors Kl K t K l specific to each satellite Sk, resp. Sb

8. Localization method according to claim 7, wherein the system of equations is solved by choosing pairs of satellites for distinct equations having line-of-sight difference vectors (==- \ and y \ orthogonal - LOSsJ LvSsk2 - LOSsJ to each other.

9. A localization method according to claim 7 or 8, wherein the local orientation of the track segments on which the vehicle is likely to be located at t is assumed to be known, and the system of equations is solved by implementing a preliminary step comprising: - selection of pair(s) of satellite(s) (Sk, Sf) having a line-of-sight difference vector / [LOSSk-LOS s,) orthogonal to said orientation and determination of the value of A ^ycommc equal to said calculated delay difference X(Sh B, Vf-X^, Bk t).

10. Computer program, intended to be stored in the memory of an electronic control device and further comprising a microcomputer, said computer program comprising instructions which, when executed on the microcomputer, implement the steps of a process according to any one of the preceding claims.

11. Electronic processing unit for the localization (10) of a craft (4) adapted to move on a set of predefined trajectory(ies) (VI,V2) and comprising a satellite receiver on board the craft to receive at time t geopositioning signals from a satellite system comprising at least s satellites Sk, k=l to s and s greater than or equal to 3; the electronic processing unit (10) comprising a database (13) storing the geographic coordinates Pi of virtual beacons B; distributed along the trajectories of said set of trajectory(s); the electronic processing unit (10) being adapted to perform the following operations considering a beacon B; : it / for k = 1 to s: calculation of the delay between the signal and geostationary received by the satellite receiver at time t from the satellite Sk and the theoretical geostationary signal calculated as being received from the satellite Sk at time t at the position of the virtual beacon Bi as defined by the geographic coordinates Pi of B; in the database, said delay being named correlation delay, ^(¾ t); i2 / for the considered beacon B; and the considered time t: calculation of representative values ​​of differences between some of said delays calculated for k = 1 to s and at least one value determined as a function of one or more other(s) of said calculated correlation delays; i3 / estimation of the location of the device based on said representative values ​​of differences between said calculated delays.

Images