Method for calibrating spatial errors, and method and system for estimating the attitude of a vehicle

Abstract

A method for calibrating spatial errors induced by phase biases having a detrimental effect on the measurements of phase differences of radio signals received by three unaligned receiving antennas of a vehicle. An inter-satellite angular deviation of a pair of satellites is estimated in two different ways: on the basis of the respective positions of the vehicle and of the satellites to obtain a theoretical inter-satellite angular deviation; and on the basis of the respective directions of incidence of the satellites relative to the vehicle, which are determined from phase measurements, to obtain an estimated inter-satellite angular deviation. The space errors are estimated on the basis of said theoretical and estimated inter-satellite angular deviations. Also, a method and system for estimating the attitude of a vehicle, in particular a spacecraft.

Claims

1. A method for of estimating attitude of a craft as a function of measurements of phase differences of radio-electric signals received by at least three non-aligned reception antennas of the craft, comprising steps of: calibrating spatial errors induced by phase biases affecting measurements of the phase differences of the radio-electric signals received by said at least three non-aligned reception antennas of the craft, said phase biases depending on a direction of incidence of the radio-electric signals with respect to the craft, comprising steps of: obtaining respective positions of the craft and a pair of satellites situated at a given instant in a radio-electric field of view of the craft; calculating by at least one processor, an inter-satellite angular distance of said pair of satellites as a function of the respective positions of the craft and said pair of satellites to obtain a theoretical inter-satellite angular distance; calculating by said at least one processor, respective directions of incidence of said pair of satellites with respect to the craft as a function of phase measurements of radio-electric signals received from said pair of satellites by the three non-aligned reception antennas of the craft; calculating by said at least one processor, the inter-satellite angular distance of said pair of satellites as a function of the respective directions of incidence of said pair of satellites with respect to the craft to obtain an estimated inter-satellite angular distance; and calculating by said at least one processor, the spatial errors for the respective directions of incidence of the said pair of satellites with respect to the craft as a function of the theoretical inter-satellite angular distance and of the estimated inter-satellite angular distance determined for said pair of satellites; and compensating the spatial errors calculated during a calculation of the attitude of the craft.

2. The method as claimed in claim 1, further comprising the step of calculating the spatial errors in a form of a parametric model whose parameters are estimated as a function of theoretical inter-satellite angular distances and of estimated inter-satellite angular distances determined for pairs of satellites.

3. The method as claimed in claim 2, further comprising the step of pre-partitioning the radio-electric field of view of the craft into a plurality of distinct cones of incidence of radio-electric signals with respect to the craft; and wherein the spatial errors are considered to be the same for any direction of incidence inside same incidence cone.

4. The method as claimed in claim 3, wherein the cones of incidence of low elevation with respect to a plane determined by phase centers of the three non-aligned reception antennas of the craft are of smaller respective solid angles than those of cones of incidence of high elevation with respect to said plane.

5. The method as claimed in claim 2, further comprising the step of calculating the spatial errors in a form of a parametric model whose basis functions are spherical harmonics.

6. The method as claimed in claim 1, wherein the satellites belong to a global satellite navigation system and the three non-aligned reception antennas are antennas of receivers of said global satellite navigation system; and further comprising the step of determining the respective positions of the satellites and of the craft in accordance with navigation information included in the radio-electric signals received from said satellites.

7. The method as claimed in claim 1, further comprising the step of calculating the spatial errors in a recursive manner, a pair of satellites by a pair of satellites or a group of pairs of satellites by a group of pairs of satellites.

8. The method as claimed in claim 1, wherein Ns satellites organized into Np pairs being situated simultaneously in a radio-electric field of view of the craft; and further comprising the step of calculating the spatial errors for each of the respective directions of incidence of the Ns satellites with respect to the craft simultaneously as a function of the theoretical inter-satellite angular distances and of the estimated inter-satellite angular distances determined for said Np pairs of satellites.

9. The method as claimed in claim 1, further comprising the step of utilizing at least one of: new pairs of satellites entering the radio-electric field of view of the craft in a course of time or a same pair of satellites several times in the course of time.

10. A non-transitory computer readable medium comprising a set of program code instructions, when executed by said at least one processor, implement the steps of calculating attitude of the craft in accordance with the method as claimed in claim 1.

11. A system to calculate an attitude of a craft as a function of measurements of phase differences of radio-electric signals, comprising: at least three non-aligned reception antennas on the craft for receiving the radio-electric signals; a processing unit comprising at least one processor configured to: obtain respective positions of the craft and a pair of satellites situated at a given instant in a radio-electric field of view of the craft, and calculate an inter-satellite angular distance of said pair of satellites as a function of the respective positions of the craft and said pair of satellites to obtain a theoretical inter-satellite angular distance; calculate respective directions of incidence of said pair of satellites with respect to the craft as a function of phase measurements of radio-electric signals received from said pair of satellites by the three non-aligned reception antennas of the craft, and calculate the inter-satellite angular distance of said pair of satellites as a function of the respective directions of incidence of said pair of satellites with respect to the craft to obtain a calculated inter-satellite angular distance; calculate the spatial errors for the respective directions of incidence of the said pair of satellites with respect to the craft as a function of the theoretical inter-satellite angular distance and of the calculated inter-satellite angular distance determined for said pair of satellites; and compensate the spatial errors calculated during a calculation of the attitude of the craft.

12. The attitude estimation system as claimed in claim 11, wherein the three non-aligned reception antennas are antennas of receivers of a global satellite navigation system.

13. A spacecraft comprising an attitude estimation system as claimed in claim 11.

14. The spacecraft as claimed in claim 13, further comprising a body comprising a plurality of substantially plane external faces; and wherein the three non-aligned reception antennas are arranged on same external face of said body.

Description

PRESENTATION OF THE FIGURES

(1) The invention will be better understood on reading the following description, given by way of wholly nonlimiting example while referring to the figures which represent:

(2) FIG. 1: a schematic representation of a spacecraft and of satellites in orbit around the Earth,

(3) FIG. 2: a schematic representation of a close-up view of the spacecraft of FIG. 1,

(4) FIG. 3: a chart illustrating the main steps of an exemplary implementation of a method for calibrating spatial errors,

(5) FIG. 4: a schematic representation of a close-up view of the spacecraft illustrating the expression of a direction of a satellite with respect to said spacecraft, in a frame fixed with respect to said spacecraft,

(6) FIG. 5: a schematic representation of a radioelectric field of view of the spacecraft,

(7) FIG. 6: a schematic representation of a particular partition of the radioelectric field of view of the spacecraft.

(8) In these figures, references that are identical from one figure to another designate identical or analogous elements. For the sake of clarity, the elements represented are not to scale, unless stated otherwise.

DETAILED DESCRIPTION OF EMBODIMENTS

(9) FIG. 1 schematically represents a spacecraft 10, whose attitude it is sought to estimate, and satellites 20 in orbit around the Earth.

(10) For example, the spacecraft 10 is in low orbit (LEO for “Low Earth Orbit”) around the Earth.

(11) The subsequent description deals in a nonlimiting manner with the case where the satellites 20 are satellites of a global satellite navigation system (GNSS for “Global Navigation Satellite System”), such as a GPS and/or Galileo system, etc.

(12) More particularly, GPS satellites 20 are considered in a nonlimiting manner. Such GPS satellites 20 are in circular orbit around the Earth, at an approximate altitude of 20000 kilometers.

(13) FIG. 1 also represents the respective directions of two GPS satellites 20 with respect to the spacecraft 10, which are represented by vectors u.sub.m and u.sub.n respectively. The angle θ.sub.mn between said directions u.sub.m and u.sub.n corresponds to the inter-satellite angular distance of the pair formed by said two GPS satellites 20.

(14) FIG. 2 schematically represents a close-up view of the spacecraft 10. As illustrated by this figure, the spacecraft 10 comprises a body 12 comprising several substantially plane external faces. For example, the body 12 is substantially cube shaped.

(15) The spacecraft 10 also comprises three non-aligned reception antennas, respectively A1, A2 and A3, suitable for receiving radioelectric signals transmitted by the GPS satellites 20. Nothing excludes, according to other examples, consideration of a more sizable number of reception antennas A1, A2, A3 to estimate the attitude of said spacecraft 10. However, to be able to estimate the attitude of said spacecraft 10 according to three axes, it is necessary to have at least three non-aligned reception antennas.

(16) The three reception antennas A1, A2, A3 belong to GPS receivers. These GPS receivers can be completely distinct or share certain means, such as for example calculation means suitable for determining the position of the craft 10 as a function of the radioelectric signals received on any one of said three reception antennas A1, A2, A3.

(17) Preferably, and as illustrated by FIG. 2, the three reception antennas A1, A2, A3 are arranged on one and the same external face of said body. In this way, it is possible to maximize a radioelectric field of view common to said three reception antennas. This makes it possible to maximize the number of GPS satellites 20 that are situated simultaneously in the radioelectric fields of view of said three reception antennas. Hereinafter, “radioelectric field of view of the spacecraft” designates the radioelectric field of view common to the three reception antennas A1, A2, A3.

(18) The three reception antennas A1, A2, A3 are organized into two reception bases: a first reception base B1 formed by the reception antennas A1 and A2, and a second reception base B2 formed by the reception antennas A2 and A3. It should be noted that nothing excludes, according to other examples, consideration also of a third reception base formed by the reception antennas A1 and A3.

(19) In a wholly nonlimiting manner, the case is considered where the two reception bases B1, B2 are substantially orthogonal.

(20) Throughout the context of the present patent application, “direction of incidence” of a GPS satellite 20 designates the expression of the direction of this GPS satellite with respect to the spacecraft 10 in a frame fixed with respect to said spacecraft. The directions u.sub.m and u.sub.n may indeed be expressed in an arbitrary frame; when they are expressed in a frame fixed with respect to the spacecraft 10, they are designated by v.sub.m and v.sub.n respectively. It is understood that, alone, the directions of incidence v.sub.m and v.sub.n of the GPS satellites 20 with respect to the spacecraft 10 encompass information on the orientation of the spacecraft 10 with respect to the GPS satellites 20, which information makes it possible, when combined with the respective absolute positions of the craft 10 and of the GPS satellites 20, to estimate the attitude of the craft.

(21) FIG. 2 illustrates an exemplary frame fixed with respect to the spacecraft 10, referred to as the “craft frame”, in which the directions of incidence of the various GPS satellites 20 will be expressed in the subsequent description. As illustrated by FIG. 2, the craft frame is substantially centered on the reception antenna A2, and is defined by three mutually orthogonal unit vectors x, y and z. The vector x is substantially collinear with the phase centers of the reception antennas A1, A2 of the reception base B1, and the vector y is substantially collinear with the phase centers of the reception antennas A2, A3 of the reception base B2.

(22) As indicated previously, the presence of reflecting and/or diffracting elements in proximity to the three reception antennas A1, A2, A3 implies that the radioelectric signals received on each of said three reception antennas will exhibit more or less sizable indirect components. Said indirect components are the origin of phase biases which depend on a direction of incidence of the radioelectric signals with respect to the spacecraft 10, and said phase biases introduce spatial errors into the estimation of the attitude of said spacecraft 10.

(23) FIG. 3 represents the main steps of an exemplary implementation of a method 50 for calibrating spatial errors, which are: 51 obtaining of respective positions of the spacecraft 10 and of Ns GPS satellites 20 situated at a given instant in the radioelectric field of view of the spacecraft, the Ns GPS satellites being organized into Np pairs, 52 estimation of respective inter-satellite angular distances of the Np pairs of GPS satellites 20 as a function of the respective positions of the craft 10 and of the Ns GPS satellites, so as to obtain so-called “theoretical” inter-satellite angular distances, 53 estimation of respective directions of incidence of the Ns GPS satellites 20 with respect to the spacecraft 10 as a function of phase measurements of radioelectric signals received from said GPS satellites by the three reception antennas A1, A2, A3 of the spacecraft 10, 54 estimation of the respective inter-satellite angular distances of the Np pairs of GPS satellites 20 as a function of the respective directions of incidence of the Ns GPS satellites with respect to the spacecraft 10, so as to obtain so-called “estimated” inter-satellite angular distances, 55 estimation of the spatial errors for each of the respective directions of incidence of the Ns GPS satellites 20 with respect to the spacecraft 10 as a function of said theoretical inter-satellite angular distances and said estimated inter-satellite angular distances calculated for said Np pairs of GPS satellites 20.

(24) Thus, the calibration of the spatial errors relies on the estimation of the respective inter-satellite angular distances of the Np pairs of GPS satellites 20 in two different ways so as to obtain: theoretical inter-satellite angular distances independent of the spatial errors, estimated inter-satellite angular distances affected by said spatial errors.
Obtaining of the Positions of the GPS Satellites and of the Spacecraft

(25) In a known manner, the GPS satellites 20 incorporate navigation information in the radioelectric signals that they transmit.

(26) Said navigation information comprises especially, for each radioelectric signal transmitted by a GPS satellite 20, the position of this GPS satellite 20 as well as the time of transmission of said radioelectric signal.

(27) Thus, the spacecraft 10, equipped with a GPS receiver, directly obtains the respective positions of the Ns GPS satellites 20 in said navigation information, and can determine in a conventional manner its position with the help of said navigation information.

(28) Determination of the Theoretical Inter-Satellite Angular Distances

(29) In the subsequent description, it is assumed, in order to simplify the equations, that the respective positions of the Ns GPS satellites 20 and of the spacecraft 10, as a function of which the theoretical inter-satellite angular distances are determined, are obtained without errors.

(30) In practice, it has been verified that the precision of said positions makes it possible in theory to have an error in the estimation of the inter-satellite angular distances of the order of 10e-04 degrees, this being negligible with respect to the precision aimed at in estimating the attitude of the spacecraft 10, which is of the order of 0.1 degrees.

(31) As illustrated by FIG. 2, the inter-satellite angular distance between two satellites of directions respectively u.sub.m and u.sub.n corresponds to an angle θ.sub.mn between said directions u.sub.m and u.sub.n.

(32) For example, the directions u.sub.m and u.sub.n are estimated as a function of the positions of the GPS satellites 20 considered and of the spacecraft 10. In the subsequent description, in a nonlimiting manner, the case is considered where the inter-satellite angular distance θ.sub.mn is estimated in the form of the scalar product u.sub.m.Math.u.sub.n, thus amounting in practice to estimating cos(θ.sub.mn) when the directions u.sub.m and u.sub.n are unit vectors.

(33) The respective positions of the spacecraft 10 and of the GPS satellites 20 are for example obtained in a fixed frame centered on the Earth. In this fixed frame centered on the Earth, we designate by: (x.sub.R, y.sub.R, z.sub.R) the coordinates of the spacecraft 10, (x.sub.m, y.sub.m, z.sub.m) the coordinates of the GPS satellite 20 of direction u.sub.m, (x.sub.n, y.sub.n, z.sub.n) the coordinates of the GPS satellite 20 of direction u.sub.n.

(34) In this case, the directions u.sub.m and u.sub.n are for example determined according to the following expressions:

(35) $u_m=\frac{1}{P_m}.Math.\left(x_m-x_R,y_m-y_R,z_m-z_R\right)$$u_n=\frac{1}{P_n}.Math.\left(x_n-x_R,y_n-y_R,z_n-z_R\right)$
in which expressions P.sub.m and P.sub.n are normalization coefficients.

(36) Next the theoretical inter-satellite angular distance between the GPS satellites 20 of directions u.sub.m and u.sub.n is determined in the form of the scalar product u.sub.m.Math.u.sub.n which is equal to:

(37) $u_m.Math.u_n=\frac{\left(x_m-x_R\right).Math.\left(x_n-x_R\right)+\left(y_m-y_R\right).Math.\left(y_n-y_R\right)+\left(z_m-z_R\right).Math.\left(z_n-z_R\right)}{P_m.Math.P_n}$
Estimation of the Directions of Incidence as a Function of Phase Measurements

(38) As indicated previously, “estimating the direction of incidence” of a GPS satellite 20 is intended to mean the act of estimating the direction of this GPS satellite with respect to the spacecraft 10 in a frame fixed with respect to said spacecraft, such as the craft frame.

(39) FIG. 4 represents in a more detailed manner the direction of incidence v.sub.m in the craft frame, so as to explain how said direction of incidence v.sub.m may be estimated.

(40) As illustrated by FIG. 4, the direction of incidence v.sub.m makes an angle α1.sub.m with the vector x of the craft frame, and an angle α2.sub.m with the vector y of said craft frame.

(41) Furthermore, in a wholly nonlimiting manner, the case is considered where the two reception bases B1, B2 are of the same length b (not illustrated by the figures).

(42) Consequently, we have cos(α1.sub.m)=Δr1.sub.m/b and cos(α2.sub.m)=Δr2.sub.m/b, in which expressions: Δr1.sub.m is the path difference of the radioelectric signals received in the direction of incidence v.sub.m by the reception antennas A1, A2 of the reception base B1, Δr2.sub.m is the path difference of the radioelectric signals received in the direction of incidence v.sub.m by the reception antennas A2, A3 of the reception base B2.

(43) It should be noted that, in these expressions, Δr1.sub.m (respectively Δr2.sub.m) is negative when α1.sub.m (respectively α2.sub.m) is greater than π/2.

(44) The direction of incidence v.sub.m is expressed for example:

(45) $v_m=\left(\frac{\Delta r{1}_m}{b},\frac{\Delta r{2}_m}{b},\sqrt{1-{\left(\frac{\Delta r{1}_m}{b}\right)}^2-{\left(\frac{\Delta r{2}_m}{b}\right)}^2}\right)$

(46) In this expression, the length b is known.

(47) Furthermore, and in a manner known to the person skilled in the art, it is possible to determine a path difference of the radioelectric signals received by the two reception antennas of a reception base B1, B2 as a function of the difference of the phases measured on said two reception antennas.

(48) It should be noted that an ambiguity may occur due to the fact that, the phase being measured modulo 2π, the path difference is estimated modulo the wavelength of the radioelectric signals received. However, methods exist, considered to be known to the person skilled in the art, for removing the ambiguity in the path difference (see for example FR 2926891).

(49) Consequently, an estimation ve.sub.m of the direction of incidence v.sub.m can be obtained by estimating the path differences Δr1.sub.m and Δr2.sub.m by methods considered to be known to the person skilled in the art.

(50) Determination of the estimated inter-satellite angular distances

(51) The estimated inter-satellite angular distances are determined as a function of the estimations ve.sub.k (1≦k≦Ns) of the directions of incidence v.sub.k of the Ns GPS satellites 20 with respect to the spacecraft 10.

(52) For example, estimations ve.sub.m and ve.sub.n of the directions of incidence respectively v.sub.m and v.sub.n are determined, for example such as described previously, and the corresponding estimated inter-satellite angular distance is determined in the form of the scalar product ve.sub.m.Math.ve.sub.n.

(53) Estimation of the Spatial Errors

(54) In the subsequent description, the case is considered where the spatial errors are estimated, for each reception base B1, B2, in the form of an error in a path difference of the radioelectric signals.

(55) Indeed, although the phase biases vary from one reception antenna to another, it is in fact the phase difference between two reception antennas which is used to estimate the path difference, itself used to estimate the attitude of the spacecraft 10. Thus, the spatial errors induced by said phase biases can actually be estimated, for each reception base B1, B2, in the form of an error in a path difference or in a phase difference between the radioelectric signals received by the two reception antennas of the reception base considered.

(56) An exemplary linear relation that may be used to estimate the spatial errors as a function of the theoretical inter-satellite angular distances and of the estimated inter-satellite angular distances is established hereinafter.

(57) More particularly, a pair of GPS satellites 20 of respective directions u.sub.m and u.sub.n (directions of incidence v.sub.m and v.sub.n in the craft frame) is considered in this example, and a linear relation is established between the following parameters: a difference ΔY.sub.mn between the theoretical inter-satellite angular distance and the estimated inter-satellite angular distance, determined for said pair of satellites, stated otherwise:
ΔY.sub.mn=ve.sub.m.Math.ve.sub.n.Math.u.sub.m.Math.u.sub.n=ve.sub.m.Math.ve.sub.n−V.sub.m.Math.v.sub.n a vector δ.sub.mn comprising the spatial errors for each reception base B1, B2 and each direction of incidence v.sub.m and v.sub.n.

(58) More particularly, this example considers the case where the vector δ.sub.mn is expressed in the following form:
δ.sub.mn=(δr.sub.m.sup.B2δr.sub.m.sup.B1δr.sub.n.sup.B2δr.sub.n.sup.B1).sup.T
in which expression: δr.sup.B2.sub.m is the error in the path difference for the reception base B2 in the direction of incidence v.sub.m, δr.sup.B1.sub.m is the error in the path difference for the reception base B1 in the direction of incidence v.sub.m, δr.sup.B2.sub.n is the error in the path difference for the reception base B2 in the direction of incidence v.sub.n, δr.sup.B1.sub.n is the error in the path difference for the reception base B1 in the direction of incidence v.sub.n.
It is noted at this juncture that: an error solely in the path difference of the reception base B2 (that is to say δr.sup.B1.sub.k=0) corresponds to a rotation about the axis of the craft frame of vector x, an error solely in the path difference of the reception base B1 (that is to say δr.sup.B2.sub.k=0) corresponds to a rotation about the axis of the craft frame of vector y.

(59) Consequently, it is possible to express the directions of incidence v.sub.m and v.sub.n as being obtained, with the help of their estimations ve.sub.m and ve.sub.n respectively, by successive rotations about the axes of the craft frame of vectors x and y.

(60) For the direction of incidence v.sub.k and its estimate ve.sub.k, k equal to m or n, this amounts to having v.sub.k=M.sub.k.Math.ve.sub.k, with:

(61) $M_k=\left(\begin{array}{ccc}1& 0& 0\\ 0& 1& -\mathrm{.Math.}{x}_k\\ 0& \mathrm{.Math.}{x}_k& 1\end{array}\right).Math.\left(\begin{array}{ccc}1& 0& \mathrm{.Math.}{y}_k\\ 0& 1& 0\\ -\mathrm{.Math.}{y}_k& 0& 1\end{array}\right)$
in which expression εx.sub.k and εy.sub.k (k equal to m or n) correspond to the angles of rotation about the axes of the craft frame of vectors x and y respectively, which angles of rotation are considered to be small so that cos(εx.sub.k) and cos(εy.sub.k) are substantially equal to εx.sub.k and εy.sub.k respectively, and so that sin(εx.sub.k) and sin(εy.sub.k) are substantially equal to 1.

(62) It is also possible to neglect the second-order terms so as to obtain the following relation:

(63) $M_k\approx \left(\begin{array}{ccc}1& 0& \mathrm{.Math.}{y}_k\\ 0& 1& -\mathrm{.Math.}{x}_k\\ -\mathrm{.Math.}{y}_k& \mathrm{.Math.}{x}_k& 1\end{array}\right)$

(64) By replacing, in the expression for the difference ΔY.sub.mn, the directions of incidence v.sub.m and v.sub.n by respectively M.sub.m.Math.ve.sub.m and M.sub.n.Math.ve.sub.n, a relation of the following type is obtained:

(65) $\Delta Y_{\mathrm{mn}}={\left(\begin{array}{c}{\mathrm{ve}}_{\mathrm{mx}}.Math.{\mathrm{ve}}_{\mathrm{ny}}-{\mathrm{ve}}_{\mathrm{my}}.Math.{\mathrm{ve}}_{\mathrm{nz}}\\ -{\mathrm{ve}}_{\mathrm{mz}}.Math.{\mathrm{ve}}_{\mathrm{nx}}+{\mathrm{ve}}_{\mathrm{mx}}.Math.{\mathrm{ve}}_{\mathrm{nz}}\\ -{\mathrm{ve}}_{\mathrm{mz}}.Math.{\mathrm{ve}}_{\mathrm{ny}}+{\mathrm{ve}}_{\mathrm{my}}.Math.{\mathrm{ve}}_{\mathrm{nz}}\\ {\mathrm{ve}}_{\mathrm{mz}}.Math.{\mathrm{ve}}_{\mathrm{nx}}-{\mathrm{ve}}_{\mathrm{mx}}.Math.{\mathrm{ve}}_{\mathrm{nz}}\end{array}\right)}^T.Math.\left(\begin{array}{c}\mathrm{.Math.}{x}_m\\ \mathrm{.Math.}{y}_m\\ \mathrm{.Math.}{x}_n\\ \mathrm{.Math.}{y}_n\end{array}\right)+O\left(\mathrm{.Math.}\right)=a\left({\mathrm{ve}}_m,{\mathrm{ve}}_n\right).Math.\left(\begin{array}{c}\mathrm{.Math.}{x}_m\\ \mathrm{.Math.}{y}_m\\ \mathrm{.Math.}{x}_n\\ \mathrm{.Math.}{y}_n\end{array}\right)+O\left(\mathrm{.Math.}\right)$
in which expression o(ε) corresponds to the second-order terms. Neglecting these second-order terms, we obtain:

(66) $\Delta Y_{\mathrm{mn}}\approx a\left({\mathrm{ve}}_m,{\mathrm{ve}}_n\right).Math.\left(\begin{array}{c}\mathrm{.Math.}{x}_m\\ \mathrm{.Math.}{y}_m\\ \mathrm{.Math.}{x}_n\\ \mathrm{.Math.}{y}_n\end{array}\right)$

(67) For the reception base B2, if Δre2.sub.k denotes the estimate of the path difference Δr2k, we have:
δr.sup.B2.sub.k=Δre2.sub.k−Δr2.sub.k=b.Math.ve.sub.k.Math.y−b.Math.v.sub.k.Math.y

(68) Because the error δr.sup.B2.sub.k does not depend on the angle of rotation εy.sub.k, we can write:

(69) $\delta r_k^{B2}=\left({\mathrm{ve}}_k-\left(\begin{array}{ccc}1& 0& 0\\ 0& 1& -\mathrm{.Math.}{x}_k\\ 0& \mathrm{.Math.}{x}_k& 1\end{array}\right).Math.{\mathrm{ve}}_k\right).Math.b.Math.y=\left(\begin{array}{ccc}0& 0& 0\\ 0& 0& b.Math.\mathrm{.Math.}{x}_k\\ 0& -b.Math.\mathrm{.Math.}{x}_k& 0\end{array}\right).Math.{\mathrm{ve}}_k.Math.y$

(70) Consequently, we obtain:
δr.sup.B2.sub.k=b.Math.εx.sub.k.Math.ve.sub.kz.

(71) In an analogous manner, the following expression is obtained for δr.sup.B1.sub.k:
δr.sup.B1.sub.k=−b.Math.εy.sub.k.Math.ve.sub.kz.

(72) In accordance with the expression hereinabove for the difference ΔY.sub.mn, we have the following relation:

(73) $\Delta Y_{\mathrm{mn}}=\frac{1}{b}.Math.{\left(\begin{array}{c}{\mathrm{ve}}_{\mathrm{mx}}.Math.{\mathrm{ve}}_{\mathrm{ny}}-{\mathrm{ve}}_{\mathrm{my}}.Math.{\mathrm{ve}}_{\mathrm{nz}}\\ -{\mathrm{ve}}_{\mathrm{mz}}.Math.{\mathrm{ve}}_{\mathrm{nx}}+{\mathrm{ve}}_{\mathrm{mx}}.Math.{\mathrm{ve}}_{\mathrm{nz}}\\ -{\mathrm{ve}}_{\mathrm{mz}}.Math.{\mathrm{ve}}_{\mathrm{ny}}+{\mathrm{ve}}_{\mathrm{my}}.Math.{\mathrm{ve}}_{\mathrm{nz}}\\ {\mathrm{ve}}_{\mathrm{mz}}.Math.{\mathrm{ve}}_{\mathrm{nx}}-{\mathrm{ve}}_{\mathrm{mx}}.Math.{\mathrm{ve}}_{\mathrm{nz}}\end{array}\right)}^T.Math.\left(\begin{array}{c}\frac{\delta r_m^{B2}}{{\mathrm{ve}}_{\mathrm{mz}}}\\ -\frac{\delta r_m^{B1}}{{\mathrm{ve}}_{\mathrm{mz}}}\\ \frac{\delta r_n^{B2}}{{\mathrm{ve}}_{\mathrm{nz}}}\\ -\frac{\delta r_n^{B1}}{{\mathrm{ve}}_{\mathrm{nz}}}\end{array}\right)$
that is to say:

(74) 0$\Delta Y_{\mathrm{mn}}=\frac{1}{b}.Math.{\left(\begin{array}{c}\frac{{\mathrm{ve}}_{\mathrm{mx}}.Math.{\mathrm{ve}}_{\mathrm{ny}}-{\mathrm{ve}}_{\mathrm{my}}.Math.{\mathrm{ve}}_{\mathrm{nz}}}{{\mathrm{ve}}_{\mathrm{mx}}}\\ \frac{{\mathrm{ve}}_{\mathrm{mx}}.Math.{\mathrm{ve}}_{\mathrm{nx}}-{\mathrm{ve}}_{\mathrm{mx}}.Math.{\mathrm{ve}}_{\mathrm{nz}}}{{\mathrm{ve}}_{\mathrm{mx}}}\\ \frac{-{\mathrm{ve}}_{\mathrm{mz}}.Math.{\mathrm{ve}}_{\mathrm{ny}}+{\mathrm{ve}}_{\mathrm{my}}.Math.{\mathrm{ve}}_{\mathrm{nz}}}{{\mathrm{ve}}_{\mathrm{nz}}}\\ \frac{-{\mathrm{ve}}_{\mathrm{mz}}.Math.{\mathrm{ve}}_{\mathrm{nz}}+{\mathrm{ve}}_{\mathrm{mx}}.Math.{\mathrm{ve}}_{\mathrm{nz}}}{{\mathrm{ve}}_{\mathrm{nz}}}\end{array}\right)}^T.Math.\left(\begin{array}{c}\delta r_m^{B2}\\ \delta r_m^{B1}\\ \delta r_n^{B2}\\ \delta r_n^{B1}\end{array}\right)$

(75) On the basis of the above relation, given by way of nonlimiting example, it is understood that any algorithm known to the person skilled in the art can be implemented to estimate the spatial errors with the help of the difference between on the one hand the theoretical inter-satellite angular distances and, on the other hand, the estimated inter-satellite angular distances.

(76) The above relation links a scalar observation (difference ΔY.sub.mn) with four parameters to be estimated (spatial errors δr.sup.B2.sub.m, δr.sup.B1.sub.m, δr.sup.B2.sub.n, δr.sup.B1.sub.n). It may for example be implemented in a recursive estimation algorithm (recursive least squares, Kalman filter, etc.). In this case, it is possible to execute step 55 of estimating the spatial errors pair of GPS satellites 20 by pair of GPS satellites (or to consider different groups of pairs of satellites at each iteration).

(77) It should be noted that, in the course of time, GPS satellites 20 may enter or exit the radioelectric field of view of the craft 10. Thus, as and when GPS satellites 20 enter the radioelectric field of view of the craft 10, they can be considered among pairs for the calibration of the spatial errors for the directions of incidence of these GPS satellites 20.

(78) Furthermore, the directions of incidence of one and the same pair of GPS satellites 20 situated in the radioelectric field of view of the craft 10 will vary in the course of time, on account of the transit of the craft 10 in LEO orbit. One and the same pair of GPS satellites 20 may therefore be considered several times in the course of time to estimate the spatial errors for different directions of incidence.

(79) More generally, it is therefore understood that the calibration method 50 carries out advantageously, in particular modes of implementation, a temporal filtering to estimate the spatial errors as a function not only of the observations at a given instant, but also as a function of earlier observations. Such temporal filtering is inherent in recursive estimation algorithms especially. Several implementations are possible for taking into account such observations carried out at different instants, and it is understood that the choice of a particular implementation merely constitutes one variant of the invention among others. According to a nonlimiting example, the steps of the method 50 for calibrating spatial errors can be iterated each time that a new pair of GPS satellites 20 has been detected. Preferably, in this case, step 55 of estimating the spatial errors implements a Kalman filter.

(80) It should be noted that nothing excludes the estimation of spatial errors for each direction of incidence of the GPS satellites 20 situated, over time and in tandem with the respective transits of said GPS satellites and of the spacecraft above the surface of the Earth, in the radioelectric field of view of the spacecraft 10.

(81) In particular modes of implementation, the spatial errors are estimated in the form of a parametric model whose parameters are estimated as a function of theoretical inter-satellite angular distances and of estimated inter-satellite angular distances calculated for various pairs of GPS satellites 20.

(82) In a preferred exemplary implementation, the spatial errors are, in the parametric model, piecewise constant. Stated otherwise, the radioelectric field of view of the spacecraft 10 is previously partitioned into a plurality of distinct cones of incidence of radioelectric signals with respect to the craft 10, and the spatial errors are considered to be the same for any direction of incidence inside one and the same incidence cone.

(83) In this way, it is possible to estimate just one spatial error per reception base B1, B2 for each incidence cone, thereby making it possible to reduce the number of calculations to be performed and/or the quantity of spatial errors to be stored.

(84) FIG. 5 schematically represents the radioelectric field of view (designated by “FV”) of the spacecraft 10 as being substantially semi-hemispherical above a plane determined by the three reception antennas A1, A2, A3. FIG. 5 also represents the convention adopted in a nonlimiting manner to express a direction of incidence v.sub.m in the form of two angles of incidence: an angle φ of azimuth, chosen as being the angle between the vector x and the projection of the direction of incidence v.sub.m on the plane determined by the three reception antennas A1, A2, A3, an angle θ of elevation, chosen as being the angle between the direction of incidence v.sub.m and the plane determined by said three reception antennas A1, A2, A3.

(85) It is for example possible to define a set of adjacent incidence cones, each of said incidence cones corresponding to a width of 5 degrees in azimuth and 5 degrees in elevation.

(86) In a particularly advantageous partition of the radioelectric field of view FV of the spacecraft 10, the cones of incidence of low elevation with respect to a plane determined by the three reception antennas A1, A2, A3 of the spacecraft 10 are of smaller respective solid angles than those of cones of incidence of high elevation with respect to said plane.

(87) It has indeed been noted that the indirect component is weaker for substantially normal directions of incidence (high elevation), with respect to the plane determined by the three reception antennas, than for the grazing directions of incidence (low elevation) with respect to said plane.

(88) Such provisions make it possible to have a smaller spatial sampling interval in the zones where the spatial errors vary a lot than in the zones where said spatial errors vary little.

(89) FIG. 6 schematically represents an example of such a radioelectric field of view FV, in which the smaller the elevation 8, the smaller the width of the incidence cones, both in azimuth and in elevation.

(90) It should be noted that nothing excludes, according to other examples, consideration of other types of parametric models suitable for modeling the spatial errors with the help of a finite number of parameters. According to a nonlimiting example, the spatial errors are estimated in the form of a parametric model whose basis functions are spherical harmonics.

(91) The method 50 for calibrating spatial errors is advantageously used in a method of estimating attitude of the spacecraft 10. For example, the spatial errors are calibrated in the form of errors in a path difference of the radioelectric signals for each reception base B1, B2. The estimation of attitude is performed for example by estimating path differences as a function of phase measurements on each of the reception antennas A1, A2, A3. Thereafter, the spatial errors are compensated on said estimated path differences, and the attitude of the spacecraft 10 is estimated as a function of the path differences obtained after compensation according to any attitude estimation algorithm known to the person skilled in the art, for example by means of a Kalman filter.

(92) The present invention also relates to a system for estimating attitude of the spacecraft 10 as a function of phase measurements of radioelectric signals received on the three reception antennas A1, A2, A3.

(93) The attitude estimation system comprises, especially, in addition to conventional means, means configured to calibrate the spatial errors in accordance with the invention. As shown in FIG. 2, these means take for example the form of a processing unit 13 comprising at least one processor 14 and an electronic memory 15 in which a computer program product is stored, in the form of a set of program code instructions to be executed by the processor so as to calibrate the spatial errors and to estimate the attitude of the craft 10. In a variant, the processing unit of the attitude estimation system comprises programmable logic circuits, of FPGA, PLD type, etc., and/or specific integrated circuits (ASICs), configured to perform all or part of the calibration of the spatial errors.

(94) It should be noted that the attitude estimation system is either distributed between the spacecraft 10 and a ground station, or embedded entirely onboard said spacecraft 10.

(95) It was considered hereinabove that the Ns satellites were satellites of a global satellite navigation system. Nothing excludes, according to other examples, consideration of other types of satellites. In such a case, the respective positions of the Ns satellites and of the craft can be obtained by any means known to the person skilled in the art. For example, said positions can be estimated by a ground station and sent to the craft for estimation of the theoretical inter-satellite angular distances.

(96) Furthermore, a spacecraft 10 was considered hereinabove. Nothing excludes, according to other examples, consideration of an aerial craft or of any object for which an estimation of orientation could be advantageous.

(97) Moreover, the invention has been described by considering radioelectric signals transmitted by satellites. According to other examples, nothing excludes consideration of aerial or terrestrial transmitters, in addition to said satellites or instead of said satellites. For example, it is possible to consider radioelectric signals transmitted by terrestrial pseudolites.

(98) The description hereinabove clearly illustrates that through its various characteristics and their advantages, the present invention achieves the objectives that it set out to attain. In particular, the present invention makes it possible to calibrate, without requiring an attitude tracker other than the reception antennas, the spatial errors induced by phase biases affecting phase measurements of radioelectric signals received by said reception antennas.