METHOD AND UNIT FOR CALCULATING INTERITIVITY NAVIGATION DATA

Description

technical field

[0001] The present invention relates to inertial navigation systems and more particularly to a method and a data processing unit for inertial navigation. Prior art

[0002] Inertial navigation systems are, for example, but not exclusively, intended to be used for the geolocation of aircraft, boats, land vehicles, ... for the control of firing, or the designation of targets.

[0003] Much work has been done to improve the accuracy of inertial navigation systems.

[0004] The improvement work is based primarily on improving the measurement accuracy of inertial sensors, such as accelerometers and gyroscopes, or on the fusion or hybridization of inertial data with data from other sensors or with complementary behavioral models.

[0005] The fusion or hybridization of inertial data and data from other sensors often makes it possible to significantly improve the typical geolocation accuracy of an inertial navigation system.

[0006] On the other hand, data hybridization often does not allow for the calculation of a reliable statistical indicator of geolocation error accuracy.

[0007] For such an indicator of geolocation error accuracy to be reliable, it is necessary to know the error statistics of the sensors considered.

[0008] This is generally the case for inertial sensors, whose manufacturers can guarantee error statistics in production. However, this is not necessarily the case for other sensors used in data hybridization.

[0009] Indeed, these data fusion techniques often rely on behavioral models of the vehicle (in this case, the aircraft), the observed scene, or limited databases, which makes it impossible to associate reliable error statistics with this type of model. Furthermore, while effective, data fusion techniques are generally insecure and cannot guarantee error statistics in cases of malfunctions that are not due to sensor failures, but rather to unrobust use—that is, outside of safe operating ranges.

[0010] In these cases, it is not enough to combine several redundant inertial measurement units because it is a common malfunction in the vehicle's integrated navigation system.

[0011] Furthermore, geolocation data from navigation systems can be used in critical functional chains, for example for the autonomous driving of a vehicle, such as a drone, an aircraft, a train, a motor vehicle, a boat, a submarine, ..., or for the control of a cannon shot or the designation of a target in a military use case.

[0012] In these cases, it is necessary to be able to certify the navigation center with a sufficient level of criticality in the face of the associated safety risk, and to be able to provide a statistical limit of geolocation data with a probability of having a navigation error outside this limit, which is compatible with the safety requirements of the mission.

[0013] In the state of the art, and particularly in the field of aeronautics, we know of inertial navigation systems using a hybrid approach with radio navigation and ensuring the fusion of inertial data with geolocation data from GPS, GLONASS, ... systems.

[0014] Error statistics are defined in the GLOBAL POSITIONING SYSTEM STANDARD POSITIONING SERVICE PERFORMANCE STANDARD [DoD & GPS NavSTAR - 2020] document, with regard to GPS, for example, and are used to define position error protection radii for GPS hybrid inertial navigation, even in the event of failure of one or more satellites.

[0015] Reference can also be made to US document 5, 760, 737 or to document WO 2012 031940 which describe examples of applications of this hybrid technique with radio navigation.

[0016] But these techniques for calculating the protection radius for position errors are only valid when used with a GPS global positioning system and do not address issues such as multipath error on the ground, spoofing, or even use cases without a GNSS signal (permanent jamming, ...). Description of the invention

[0017] The aim of the invention is therefore to overcome the aforementioned disadvantages and to provide a method and a central processing unit for inertial navigation data that is capable of delivering certified geolocation data by providing geolocation data with an accuracy indicator associated with a probability of error.

[0018] Another objective of the invention is to use data from uncertified hybrid inertial navigation in a critical inertial navigation functional chain.

[0019] The invention therefore relates, according to a first aspect, to a method for calculating inertial navigation data according to claim, in which: We acquire navigation data, and we calculate hybrid navigation values ​​from this navigation data.

[0020] According to this process, certified navigation values ​​are calculated, independently of hybrid navigation values, the said certified navigation values ​​being certified with a geolocation error limit.

[0021] For certified navigation values, we calculate a first permissible geolocation error limit corresponding to geolocation with a predetermined probability of errors.

[0022] The invention thus enables the calculation, in parallel with a high-performance but uncertified hybrid inertial navigation system without a reliable error protection limit, based on the hybridization of inertial data, of a simpler, certified inertial navigation system. This simpler system allows for the calculation of a reliable error protection limit for the hybrid inertial navigation data, without requiring certification of the hybrid navigation system itself. The error protection limit is defined as a statistical limit on the geolocation data, with a probability of a navigation error occurring outside this limit.

[0023] Hybrid navigation values ​​are certified by calculating a protection limit based on a comparison with a secure and certified inertial navigation, unaffected by external sensors or inconsistent behavioral mathematical models.

[0024] Thus, hybrid navigation values ​​can be integrated and exploited in a certified critical functional chain.

[0025] For hybrid navigation values, a second permissible geolocation error limit is calculated from the first geolocation error limit and a difference between the hybrid geolocation values ​​and the certified geolocation values.

[0026] According to another feature of the method according to the invention, a validity status of hybrid navigation values ​​is calculated by comparison with respective threshold values.

[0027] In one implementation mode, when calculating certified navigation values, a Kalman filter is used to calculate error covariances from the navigation data.

[0028] For example, the Kalman filter is an invariant Kalman filter.

[0029] According to yet another feature of the method according to the invention, a zero-speed recalibration is carried out on an inertial navigation stage capable of calculating certified navigation values.

[0030] It can be predicted that hybrid navigation values ​​will be calculated from inertial data from sensors, including three gyroscopes and three accelerometers, in three directions of space, and additional data, including displacement models.

[0031] The invention also relates to an inertial navigation system according to claim 7, comprising a navigation data acquisition unit and a first navigation stage capable of calculating hybrid navigation values ​​by fusion of navigation data.

[0032] This control unit also includes a second navigation stage capable of calculating certified navigation values, independently of hybrid navigation values, said certified navigation values ​​being certified with a geolocation error limit. Brief description of the drawings

[0033] Other objects, features and advantages of the invention will become apparent from the following description, given by way of non-limiting example and with reference to the accompanying drawings in which: [ Fig 1 ] is a diagram of the hardware architecture of an inertial navigation system according to the invention; [ Fig 3 ] illustrates the main steps of a method for calculating inertial navigation data according to the invention [ Fig 3[ ] shows a diagram illustrating the first and second navigation data with their permissible geolocation error limit in the form of position error limit protection circles; and A detailed description of at least one embodiment of the invention

[0034] We refer first to the figure 1 which schematically illustrates the architecture of an inertial navigation system according to the invention.

[0035] This control unit comprises an inertial navigation data acquisition unit (IMU) 1, including measurement sensors, for example, three gyroscopes and three accelerometers providing inertial data in three spatial directions; a first hybrid navigation stage 2, designed to calculate hybrid inertial navigation values ​​from the inertial navigation data provided by the acquisition unit 1 and from complementary sensors or behavioral models 4, which provide additional navigation data, such as GPS geolocation data, distances traveled, a vehicle movement model, etc.; and a second navigation stage 3, providing safe navigation values. The first and second inertial navigation stages each consist of a computer programmed to implement the functions described below.These calculators can, however, be formed by calculation partitions of the same calculator.

[0036] The first and second computing stages accurately calculate navigation data values

[0037] The first inertial navigation stage 2 ensures the calculation of hybrid navigation data, for example position, speed and altitude, in three dimensions by hybridization, i.e. by merging the data delivered by the inertial data acquisition unit 1 and the external help data provided by the sensors or the behavioral models 4.

[0038] This first computing stage includes a computer 5, which performs the calculation of this hybrid navigation data. As will be described in detail later, it also receives position recalibration and initialization data. P .

[0039] The second computing stage comprises a computer 6, which performs the calculation of certified navigation data independently of the hybrid navigation data provided by the acquisition unit 1, and an additional computing module 7 for calculating geolocation error protection limits. Computer 6 receives as input position adjustment, zero-speed recalibration, and initialization data. P'.

[0040] Hybrid navigation calculation is a function that primarily calculates an inertial geolocation solution (3D position, 3D velocity, 3D attitudes) using the fusion of UMI measurements and data from external aids or behavioral models of the carrier (radio navigation receiver, odometer and vehicle movement model, log sensor (for a marine carrier), zero-speed recalibration (ZUPT), optical sensor, lidar, etc.) for example using techniques such as the integration of inertial navigation coupled with a Kalman filter which allows the calculation of error covariances from the characteristics of the inertial sensors and the aid sensors or using other techniques such as particle filtering or neural networks and artificial intelligence.

[0041] The hybrid navigation calculation is a function that secondly estimates the 3x3 rotation matrix for transitioning from the local geographic reference frame "North-East-Down" [N] to the "Body" reference frame [B] linked to the vehicle: C N B ^ hybrid the notation ^ denoting an estimate of the magnitude (), ^ therefore being tainted by error

[0042] This is yet another function that allows for the estimation of the vehicle's velocity components. V N ^ ¯ hybrid in the local geographic coordinate system "North-East-Down" [N], the estimated horizontal position coordinates of the vehicle are expressed according to the WGS84 ellipsoid model, for example the estimated longitude Ĝ hybrid and estimated latitude L hybrid, the estimate ĥ hybrid of the vehicle's altitude expressed relative to the geoid.

[0043] The hybrid navigation calculation also ensures the calculation of a covariance matrix for geolocation errors if a Kalman filter is used, based on the modeling of geolocation errors. Geolocation errors for hybrid navigation are further defined through mathematical modeling.

[0044] This includes, for example, angle errors, in 3 dimensions, of the matrix C N B ^ hybrid the following, projected onto the Body [B] coordinate system: φ B ¯ hybrid = φ X B hybrid φ Y B hybrid φ Z B hybrid , defined by the following relationships: φ B ¯ hybrid = logSO 3 C N B . C N B ^ hybrid T expSO 3 φ B ¯ hybrid × = C N B . C N B ^ hybrid T With Vect B ¯ × = 0 − Vect Z B Vect Y B Vect Z B 0 − Vect X B − Vect Y B Vect X B 0 the 3x3 antisymmetric matrix of the vector Vector B< expressed in the frame [B]; And where: "logSO3" is the logarithmic map of the group of rotation matrices SO(3) to the rotation vectors R3 (R in the sense of the set of real numbers) and “expSO3”is the exponential map of the rotation vectors R3 of the rotation matrix group SO(3), SO(3) being the special orthogonal group of 3-dimensional square matrices

[0045] The error estimates of three-dimensional velocity values, subject to error, delivered by the first stage of hybrid navigation 2, and expressed in the local geographic frame [N] by V N ˜ ¯ hybrid = V X N ˜ hybrid V Y N ˜ hybrid V Z N ˜ hybrid are defined by the following relationship: V N ˜ ¯ hybrid = V N ¯ − V N ^ ¯ hybrid

[0046] The two-dimensional horizontal position errors expressed in the "North-East" plane of the local geographic coordinate system [N]: r X N ˜ hybrid And r Y N ˜ hybrid , are defined by the following relationships: r X N ˜ hybrid = L − L ^ hybrid ⋅ R X N L + h r Y N ˜ hybrid = G − G ^ hybrid . cos L . R Y N L + h

[0047] The radial horizontal position error can also be defined as the magnitude of these two components. r horz ˜ hybrid = r X N ˜ hybrid 2 + r Y N ˜ hybrid 2 With R XN ( L ) the Earth's radius according to the WGS 84 model in the North direction With R YN ( L ) the Earth's radius according to the WGS 84 model in the East direction. The altitude errors delivered by the first stage of hybridized inertial navigation 2, which correspond to the position error h̃ safe along the vertical axis of the local geographic coordinate system [N] are defined by h ˜ hybrid = h − h ^ hybrid

[0048] This data fusion makes it possible to improve the accuracy of geolocation by reducing the errors presented above, to estimate and compensate for errors of inertial sensors, to estimate and compensate for errors of external aids, and to estimate a statistical performance indicator, for example a covariance matrix of the errors of a Kalman filter.

[0049] With regard to the second navigation stage 3, the calculation of certified navigation, called secure, is a function allowing the calculation of an inertial geolocation solution (3D position, 3D velocity, 3D attitudes) using the fusion of measurements from the UMI 1 unit and secure external aid data, for example zero-speed recalibration data when the vehicle is stopped, or ZUPT recalibration, secured by a stop indicator transmitted to the system assumed to be at the appropriate level of operational safety; point position recalibration data or any recalibration using an integrated sensor for which a statistical probability of failure can be associated and whose errors, excluding failure, can be bound statistically and independently of the use cases considered.

[0050] These data fusion calculations are performed through the integration of inertial navigation coupled with a Kalman filter which allows the calculation of error covariances from the characteristics of inertial sensors.

[0051] The second navigation calculation stage 3 ensures the calculation of the estimation of the 3x3 rotation matrix for the transition from the local geographic reference frame "North-East-Down" [N] to the "Body" reference frame [B] linked to the vehicle: C N B ^ safe

[0052] The rating ^ denoting an estimate of the magnitude (), ^ therefore being tainted by error

[0053] It also provides the estimation of the three-dimensional components of velocity V N ^ ¯ safe of the vehicle in the local geographic coordinate system "North-East-Down" [N], as well as the estimation of the vehicle's horizontal position coordinates expressed according to the WGS84 ellipsoid model, for example, the estimated longitude Ĝsafe and estimated latitude L safe

[0054] The second stage of navigation calculation 3 also calculates the estimated altitude of the vehicle I'm safe expressed with respect to the geoid, as well as the calculation of a covariance matrix of geolocation errors provided by the Kalman filter based on the modeling of geolocation errors.

[0055] This matrix " P The 9x9 matrix "safe" represents the covariance estimates of the nine chosen geolocation errors, namely: the three-dimensional angle errors of the matrix C N B ^ safe projected onto the Body [B] coordinate system: φ B ¯ safe = φ X B safe φ Y B safe φ Z B safe , defined by φ B ¯ safe = logSO 3 C N B . C N B ^ safe T expSO 3 φ B ¯ safe × = C N B . C N B ^ safe T

[0056] Three-dimensional velocity errors expressed in the local geographic coordinate system [N]: V N ˜ ¯ safe = V X N ˜ safe V Y N ˜ safe V Z N ˜ safe , defined by V N ˜ ¯ safe = V N ¯ − V N ^ ¯ safe

[0057] The two-dimensional horizontal position errors expressed in the "North-East" plane of the local geographic coordinate system [N]: r X N ˜ safe And r Y N ˜ safe defined by: r X N ˜ safe = L − L ^ safe . R X N L + h And r Y N ˜ safe = G − G ^ safe . cos L . R Y N L + h

[0058] Altitude errors correspond to the position error along the vertical axis of the local geographic reference frame [N]: h̃ safe, as defined by h ˜ safe = h − h ^ safe

[0059] It is also possible to define the radial horizontal position error as the magnitude of these two components: r horz ˜ safe = r X N ˜ safe 2 + r Y N ˜ safe 2

[0060] With : R XN ( L ) the Earth's radius according to the WGS 84 model in the North direction, and R YN ( L ) the Earth's radius according to the WGS 84 model in the East direction

[0061] This data fusion improves geolocation accuracy by reducing the errors described above, estimates and compensates for inertial sensor errors, and allows for the estimation of a statistical performance indicator, such as a covariance matrix of Kalman filter errors. This indicator is free from errors related to poor use cases of external aids, since no complex or insecure external aids are employed. It is only affected by errors related to system failure, but this failure is known, low, and corresponds to the probability of an undetected system failure.

[0062] This geolocation system is simple, secure, yet still effective, using ZUPT or point-based positioning adjustments to provide a secure protection radius usable during operations. These point-based adjustments maintain a level of secure geolocation accuracy acceptable for the mission.

[0063] The second navigation stage 3 also calculates the protection limit for geolocation data and the validity of hybrid data. This calculation allows, in particular, the definition of a statistical limit for geolocation data, along with a probability of navigation errors occurring outside this limit.

[0064] This protection limit, denoted "PL", for errors in the secure navigation geolocation data is calculated using the "Psafe" covariance matrix of geolocation errors from the secure navigation filter (covariance matrix provided by the Kalman filter). It contains the variances of the geolocation errors and their correlations with each other.

[0065] This covariance matrix is ​​the result of propagating the error characteristics of the inertial sensors (Gaussian noise statistics, bias errors, sensor scaling factor errors, sensor misalignment errors, etc.) to the geolocation errors and taking into account zero-speed (ZUPT) or absolute position recalibrations by the navigation filter.

[0066] The statistical probability of being outside this protection limit in the absence of a failure of one of the inertial sensors, denoted "Proba", is for example: Proba = 1 e − 5

[0067] In what follows, examples of protection limit calculations will be detailed.

[0068] Firstly, the protection limit can be defined by a protection circle limiting the horizontal position error: HPL Proba = Q inv Proba 2 . VAR MAX

[0069] Or VAR max is the maximum eigenvalue of the 2x2 covariance matrix « P safe r X N ˜ safe r Y N ˜ safe r X N ˜ safe r Y N ˜ safe "Horizontal position errors North and East in the Kalman filter of secure navigation" Qi nv ( Probability, N ) is the inverse of the cumulative distribution function of the normalized Gaussian random variable of dimension N and uncorrelated for a probability < Proba. For example, for N = 2, Q inv Proba 2 = 2 . ln 1 Proba ln(x) = natural logarithm evaluated at x. For N = 2, Proba = 1e-5 and Qinv(1e-5, 2), which is approximately 4.7985

[0070] In the absence of a navigation system failure, this calculated HPL(Probability) limit is greater than r horz ˜ safe with a probability at least equal to "1-Probability"

[0071] In the absence of failure of the navigation unit, this HPL(Proba) limit protection radius thus calculated, centered on the true horizontal position of the vehicle must contain with a probability at least equal to "1-Proba" the position calculated by the secure navigation.

[0072] The protection limit can also be applied to the vertical position error (altitude): ZPL Proba = Q inv Proba 1 . P safe h ˜ safe h ˜ safe

[0073] Or P safe ( h̃ safe, h̃ safe) is the variance of the altitude error of the Kalman filter for safe navigation and Qi nv ( Probability, N) is the inverse of the cumulative distribution function of the normalized Gaussian random variable of dimension N and uncorrelated for a probability < Proba. For example, for N = 1, Q inv Proba 1 = 2 . erfcinv Proba Where erfcinv is the inverse of the complementary Gaussian error function. For N = 1, Proba = 1e-5 and Qinv(1e-5 , 1) ≈ 4.4172

[0074] In the absence of a failure of the navigation system, this ZPL(Proba) limit thus calculated is greater than | h̃ safe | with a probability at least equal to "1-Probability"

[0075] The error protection limit can also be applied to the north, east, and vertical velocity components in the local geographic coordinate system: VnPL Proba = Q inv Proba 1 . P safe V X N ˜ safe V X N ˜ safe

[0076] Or P safe V X N ˜ safe V X N ˜ safe is the variance of the velocity error along the north axis of the local geographic coordinate system of the safe navigation Kalman filter and Qi nv ( Probability, N) is the inverse of the cumulative distribution function of the normalized Gaussian random variable of dimension N and uncorrelated for a probability < Proba

[0077] In the absence of a navigation system failure, this VnPL(Proba) limit, calculated in this way, is greater than V N X ˜ safe with a probability at least equal to "1-Probability".

[0078] Similarly, for the respective East and vertical velocity errors, we have: VePL Proba = Q inv Proba 1 . P safe V Y N ˜ safe V Y N ˜ safe VzPL Proba = Q inv Proba 1 . P safe V Z N ˜ safe V Z N ˜ safe

[0079] The protection limit for three-dimensional angle errors projected onto the Body [B] frame, which is a frame of interest for these errors in a target designation case in particular, of the matrix C N B ^ safe , is written: CbxPL Proba = Q inv Proba 1 . P safe φ X B safe φ X B safe Or P safe φ X B safe φ X B safe is the variance of the Kalman filter for safe navigation of the x-axis component of the Body [B] coordinate system of angle errors of the matrix C N B ^ safe And Qi nv ( Probability, N) is the inverse of the cumulative distribution function of the normalized Gaussian random variable of dimension N and uncorrelated for a probability < Proba

[0080] In the absence of a navigation system failure, this calculated CbxPL(Proba) limit is greater than φ X B safe with a probability at least equal to "1-Probability"

[0081] Similarly, for the projected angle errors in the Body [B] frame of the matrix C N B ^ safe Along the YB and ZB axes, we have: CbyPL Proba = Q inv Proba 1 . P safe φ Y B safe φ Y B safe CbzPL Proba = Q inv Proba 1 . P safe φ Z B safe φ Z B safe

[0082] The second certified navigation stage 3 also ensures the calculation of a protection limit, noted "PL_hybrid", of errors in the geolocation data of the hybrid navigation using the PL protection limits of the errors in the geolocation data of the secure navigation associated with the probability "Proba" as described previously, and the difference between the geolocation data of the hybrid navigation and the secure navigation.

[0083] Various protection limits, which are based on calculations of the protection limits defined for secure geolocation data, can be calculated.

[0084] We can first calculate a limiting protection circle for the hybrid position error in the horizontal plane of the local geographic coordinate system: HPL_hybrid Proba = HPL Proba + Δ Pos

[0085] Or : HPL(Proba) is the limit protection circle for horizontal position errors in secure navigation presented earlier, and Δ Pos = d X N 2 + d X N 2 is the distance between the two secure and hybrid positions in the horizontal plane of the local geographic reference frame [N],

[0086] With d Y N = L ^ hybrid − L ^ safe . R X N L ^ safe + h ^ safe And d Y N = G ^ hybrid − G ^ safe . cos L ^ safe . R Y N L ^ safe + h ^ safe

[0087] In the absence of a navigation system failure, the calculated HPL_hybrid(Proba) is greater than r horz ˜ hybrid with a probability greater than or equal to "1-Probability"

[0088] Indeed, by construction (Cauchy-Schwartz inequality), r horz ˜ hybrid ≤ r horz ˜ safe + Δ Pos

[0089] However, there is a probability greater than or equal to "1-Proba", by choice of HPL(Proba), that r horz ˜ safe ≤ HPL Proba .

[0090] Therefore, we can guarantee, by choosing HPL_hybrid(Proba), that there is a probability greater than or equal to "1-Proba" that r horz ˜ hybrid ≤ HPL Proba + Δ Pos = HPL _ hybrid Proba

[0091] In the absence of a failure of the navigation unit, this protection radius limit HPL_hybrid(Proba) thus calculated, centered on the true horizontal position of the vehicle must contain with a probability at least equal to "1-Proba" the position calculated by the hybrid navigation.

[0092] Secondly, we can calculate a protection limit for the hybrid vertical position error (altitude): ZPL_hybrid Proba = ZPL Proba + Δh

[0093] Or HPL(Proba) is the limiting protection circle for horizontal position errors in secure navigation presented previously Δ h = | ĥ hybrid - h̃ safe | is the distance between the two safe and hybrid altitudes.

[0094] In the absence of a failure of the navigation system, ZPL_hybrid(Proba) thus calculated is greater than | h̃ hybrid | with a probability greater than or equal to "1-Probability"

[0095] Indeed, by construction (Cauchy-Schwartz inequality), h ˜ hybrid ≤ h ˜ safe + Δ h

[0096] However, there is a probability greater than or equal to "1-Proba", by choice of ZPL(Proba), that | h̃ safe | ≤ ZPL ( Probability ) .

[0097] We can therefore guarantee by choosing ZPL_hybrid(Proba) that there is a probability greater than or equal to "1-Proba" that | h̃ hybrid | ≤ (ZPL(Proba) + Δ h ) = ZPL_hybrid(Probability)

[0098] We can also calculate a protection limit for hybrid North, East and Vertical velocity errors in the local geographic coordinate system.

[0099] For example, using the same principle described above, we can define, for the protection limit of hybrid speed errors North, a protection limit for the error V X N ˜ hybrid , based on the relationship: VnPL_hybrid Proba = VnPL Proba + ΔVn

[0100] Or: VnHPL(Proba) is the limit of protection against velocity error along the north axis of the local geographic coordinate system [N], and Δ Vn = V X N ^ hybrid − V X N ^ safe is the difference between the two speeds: north, secure, and hybrid

[0101] In the absence of a navigation system failure, Vn_hybrid(Proba) calculated in this way is greater than V X N ˜ hybrid with a probability greater than or equal to "1-Probability"

[0102] We can also define an error protection limit V Y N ˜ hybrid : VePL_hybrid Proba = VePL Proba + ΔVe

[0103] Or : VeHPL(Proba) is the limit of protection against velocity error along the East axis of the local geographic coordinate system [N], and Δ Ve = V Y N ^ hybrid − V Y N ^ safe = écart between the two speeds is safe and hybrid

[0104] In the absence of a navigation system failure, Ve_hybrid(Proba) calculated in this way is greater than V Y N ˜ hybrid with a probability greater than or equal to "1-Probability"

[0105] We can still define a limit for error protection. V Z N ˜ hybrid : VzPL_hybrid Proba = VzPL Proba + ΔVz

[0106] Or : VzHPL(Proba) is the limit of protection against velocity error along the vertical axis of the local geographic coordinate system [N], and Δ Vz = V Z N ^ hybrid − V Z N ^ safe is the difference between the two secure and hybrid vertical speeds

[0107] In the absence of a navigation system failure, Vz_hybrid(Proba) calculated in this way is greater than V Z N ˜ hybrid with a probability greater than or equal to "1-Probability"

[0108] We can still calculate a protection limit for three-dimensional angle errors of the matrix C N B ^ hybrid projected in the Body [B] coordinate system.

[0109] For example, the error protection limit φ X B hybrid East : CbxPL _ hybrid Proba = CbxPL Proba + Δ φ X B

[0110] Where: CbxPL_hybrid(Proba) is the component protection limit according to XB of the matrix angle errors C N B ^ safe And Δφ B ¯ = Δ φ X B Δ φ Y B Δ φ Z B = logSO 3 C N B ^ hybrid . C N B ^ safe T

[0111] In the absence of a navigation system failure, the calculated CbxPL_hybrid(Proba) is greater than φ X B hybrid with a probability greater than or equal to "1-Probability"

[0112] Indeed, we have: C N B . C N B ^ hybrid T = C N B . C N B ^ safe T . C N B ^ safe ︸ identité I 3 . C N B ^ hybrid T = C N B . C N B ^ safe T . C N B ^ safe . C N B ^ hybrid T

[0113] Or again C N B . C N B ^ hybrid T = C N B . C N B ^ safe T . C N B ^ hybrid . C N B ^ safe T T

[0114] And so expSO 3 φ B ¯ hybrid × = expSO 3 φ B ¯ safe × . expSO 3 Δφ B ¯ × T expSO 3 φ B ¯ hybrid × = expSO 3 φ B ¯ safe × . expSO 3 − Δφ B ¯ ×

[0115] However, for angles of small amplitude, for example with a magnitude less than 10e-3 radians, we have, to a first order: expSO 3 φ B ¯ hybrid × ≈ I 3 + φ B ¯ hybrid × expSO 3 − Δ φ B ¯ × ≈ I 3 − Δ φ B ¯ ×

[0116] For small angles, for example with a magnitude less than 10e-3 radians, we deduce, to a first order: I 3 + φ B ¯ hybrid × = I 3 + φ B ¯ safe × − Δ φ B ¯ ×

[0117] And so φ B ¯ hybrid × = φ B ¯ safe × − Δ φ B ¯ × = φ B ¯ safe − Δ φ B ¯ ×

[0118] By the uniqueness between the antisymmetric matrix of a vector and that same vector, we deduce: φ B ¯ hybrid = φ B ¯ safe − Δφ B ¯

[0119] We then have the following inequality for the component along XB: φ X B hybrid ≤ φ X B safe + Δ φ X B

[0120] However, there is a probability greater than or equal to "1-Proba", by choice of CbxPL(Proba), that φ X B safe ≤ CbxPL Proba

[0121] Therefore, we can guarantee, by choosing CbxPL_hybrid(Proba), that there is a probability greater than or equal to "1-Proba" that φ X B hybrid ≤ CbxPL Proba + Δ φ X B = CbxPL _ hybrid Proba

[0122] Similarly, the limits of error protection φ Y B hybrid And φ Z B hybrid are calculated. For example, with regard to φ Y B hybrid This limit is calculated from the following relationship: CbyPL _ hybrid Proba = CbyL Proba + Δ φ Y B

[0123] Where: CbyPL_hybrid(Proba) is the YB component protection limit for angle errors in the matrix C N B ^ safe , And Δφ B ¯ = Δ φ X B Δ φ Y B Δ φ Z B = logSO 3 C N B ^ hybrid . C N B ^ safe T

[0124] In the absence of a navigation system failure, the calculated CbyPL_hybrid(Proba) is greater than φ Y B hybrid with a probability greater than or equal to "1-Probability".

[0125] The CbzPL_hybrid(Proba) limit is calculated in a similar way.

[0126] With reference to the figure 2 , which illustrates the main phases of a method for calculating inertial navigation data according to the invention, according to a first step 10, the inertial data acquisition unit 1 acquires data from at least three gyroscopes and at least three inertial accelerometers which measure inertial rotations and specific inertial forces in the three directions of space.

[0127] Similarly, during this step 10, data from the other 4 sensors are also acquired.

[0128] It should be noted that the sensors of the inertial data acquisition unit 1 are developed and certified according to the required safety constraints, depending on the criticality of the feared events.

[0129] During the following step 11, the first navigation stage 2 delivers the first navigation data by fusion of inertial data from the inertial data acquisition unit 4 and any other complementary navigation data from the additional sensors 4.

[0130] It should be noted that at the end of this calculation step 11, the navigation data provided at the end of this first calculation step are relatively accurate but are difficult to certify.

[0131] In the following step 12, the second navigation stage 3 ensures the calculation of secure navigation data from, in particular, error covariances and from the characteristics of inertial sensors, as described previously.

[0132] It should be noted that these acquisition and calculation steps are carried out continuously.

[0133] However, position recalibration phases are performed, for example when passing through known locations. Zero-speed recalibration phases are also implemented, particularly to recalibrate speed data.

[0134] In the following step 13, a calculation of protection limits is carried out which corresponds to a level of protection error, for the first and second stages of inertial navigation.

[0135] By also referring to the figure 3 , where P designates a real position, and P1 and P2 designate the position calculated by the first and second navigation stage 2 and 3, respectively, during this step, the error protection limit of the secure navigation geolocation data from the second inertial navigation stage is calculated, as indicated previously.

[0136] This protection limit, denoted HPL, constitutes here a protection circle limiting the geolocation error centered on the position P2 calculated by the second navigation stage 3.

[0137] In the following step 14, the error protection limit of the geolocation data calculated by the first navigation stage 2, as previously described, is calculated.

[0138] This HPL_hybrid protection limit forms a protection circle limiting the position error provided by the first navigation stage 2, centered on the position P1 geolocated by this first navigation stage and whose radius is constituted by the sum of the difference between the geolocation data provided by the first and second navigation stages and the HPL protection limit calculated during the previous step 13.

[0139] In the following step 15, a validity status V is calculated for the first hybrid navigation data by comparison with respective threshold values.

[0140] First, a consistency test of the hybrid geolocation data is carried out beforehand to ensure the consistency of the data, such as the declared validity of the data, the data format, the data interval, etc. Otherwise, an invalidity status of the data is issued.

[0141] Furthermore, the measured differences between hybrid navigation and optimal navigation calculated previously are also compared to thresholds in order to raise anomaly statuses to the user and inform them to no longer consider hybrid navigation data.

[0142] For example, a fraction of the "PL" error protection limits of secure navigation can be used as an anomaly lifting threshold.

[0143] The validity calculation can be of the following type: If the consistency test of the hybrid geolocation data is invalid, or if ΔPos > Threshold_pos, or if Δh > Threshold_Z, or if ΔVn > Threshold_Vn, etc., then the hybrid navigation data is invalid. Otherwise, the hybrid navigation data is valid. End If

[0144] It should be noted that the calculations implemented within the navigation unit are carried out within two separate computers, or partitions of a computer, respectively ensuring the calculation of hybrid navigation data and the calculation of secure navigation data, which makes it possible not to constrain the lower software and hardware layers of the hybrid navigation calculation unit which is not in the critical functional chain of the invention.

[0145] The proposed solution allows the use of data from any uncertified hybrid navigation in a critical functional chain by securing this data with a certified software or hardware partition which calculates a protection limit for this uncertified navigation data based on comparison with a secure navigation, not impacted by external sensors or inconsistent behavioral mathematical models and certified.

[0146] Thus, any navigation hybridization technique (complex Kalman filter, particle filter, neural network, artificial intelligence, ...) can be used without having to certify these processes, while guaranteeing a level of data integrity through the calculation of reliable and certified protection limits.

[0147] The data from secure navigation and their protection limits are also available as a backup solution, used for example when the difference between the geolocation data from hybrid and secure navigation is greater than a threshold, for example.

Claims

1. Method for calculating inertial navigation data, wherein: - navigation data are collected; and - hybrid navigation values are calculated from these navigation data, characterised in that: certified navigation values are calculated independently of hybrid navigation values, said certified navigation values being certified with a geolocation error limit. Wherein a first permissible geolocation error limit is calculated for the certified navigation values, corresponding to a geolocation with a predetermined probability of error, and wherein a second permissible geolocation error limit is calculated for the hybrid navigation values based on the first geolocation error limit and a deviation between the hybrid navigation values and the certified navigation values.

2. Method according to claim 1, wherein a validity status (V) of the hybrid navigation values is calculated by comparison with respective threshold values.

3. Method according to one of claims 1 and 2, wherein, when calculating the certified navigation values, a Kalman filter is used for calculating error covariances from the navigation data.

4. Method according to claim 3, wherein the Kalman filter is an invariant Kalman filter.

5. Method according to any one of claims 1 to 4, wherein a zero-speed reset of the inertial navigation stage capable of calculating the certified navigation values is performed.

6. Method according to any one of claims 1 to 5, wherein the hybrid navigation values are calculated from inertial data coming from measurement sensors, in particular from at least three gyroscopes and at least three accelerometers, in three directions in space, and additional data, particularly from movement models.

7. Inertial navigation system, comprising a navigation data acquisition unit (1) and a first navigation stage (2) capable of calculating hybrid navigation values through the fusion of navigation data, characterised in that it furthermore includes a second navigation stage (3) able to calculate certified navigation values independently of the hybrid navigation values, said certified navigation values being certified with a geolocation error limit, and in that the second navigation stage computes a first permissible geolocation error limit corresponding to a geolocation with a predetermined probability of error, and a second permissible geolocation error limit derived from the first geolocation error limit and a deviation between the hybrid navigation values and the certified navigation values.