Hybrid navigation with decoy detection via deviation monitoring

Abstract

A navigation method comprising the continuous calculation of an inertial position and, at successive time intervals, the acquisition of a satellite position to calculate a hybrid position by combining the inertial and satellite positions. The method comprises: calculating an instantaneous error at a given time and then a sustained error from the time of calculation of said instantaneous error to the current time, so as to have, for a given period of time, a plurality of sustained errors from different calculation times to the same current time; calculating, for each sustained error, at least one detection indicator; and comparing said detection indicator to a threshold to detect a failure of the hybrid position at the current time. Figure from the abstract: Fig. 1

Description

Title of the invention: Hybrid navigation with decoy detection by deviation monitoring

[0001] The present invention relates to the field of navigation and more particularly to hybrid inertial satellite navigation.

[0002] BACKGROUND OF THE INVENTION

[0003] Inertial navigation systems are known comprising an inertial measurement unit connected by a data link to an electronic navigation calculation unit.

[0004] The inertial measurement unit comprises accelerometers and angular sensors arranged along the axes of a measurement frame to provide primary signals representative of the integral, over a time step, of the specific force vector relative to a reference inertial frame. The successive signals are thus representative of the integral of the specific force vector from an instant t0 to an instant tb then from the instant ti to an instant t2, then from the instant t2 to an instant t3, etc.): the signals are therefore generally called increments. The specific force (in English "specific force", "g-force" or "mass-specific force") is, itself, a representation of the sum, on the one hand, of the acceleration of the carrier (for example a vehicle) of the inertial measurement unit relative to the inertial frame and, on the other hand, of the Earth's gravity.

[0005] The electronic navigation calculation unit comprises a processor and a memory containing a navigation computer program which is executed by the processor and which uses the signals supplied by the inertial measurement unit to determine inertial location data of the vehicle carrying the navigation system, this data comprising the position, speed and attitude of the vehicle which are calculated from the signals coming from the inertial measurement unit.

[0006] It is known that inertial location data are affected by errors linked in particular to inertial sensors, such as drift errors, biases, and scale factor errors. These errors therefore affect the accuracy of navigation.

[0007] To improve navigation accuracy, consideration has been given to using more expensive inertial sensors and optimizing electronic signal processing.

[0008] It is also known to use external location data, for example from a satellite navigation system (so-called GNSS systems, such as GPS, GALILEO, GLONASS, BEIDOU, etc.), to periodically recalibrate inertial navigation. This is known as hybrid navigation.

[0009] It is recalled that satellite positioning or localization consists of receiving signals emitted by satellites whose position is known and deducing the duration (or flight time), between the transmission and reception of each of the signals, a so-called pseudo-distance measurement separating the receiver of the satellite signals and each of the satellites from which the signal was received (each signal comprising a satellite identifier and the time of transmission of the signal). Thus, it is sufficient to have the signals from four satellites to have the latitude, longitude and altitude of the receiver, which form satellite location data, as well as an error on the measured durations, but the location is all the more precise as the number of satellites whose signals have been taken into account by the receiver to calculate its position is large.

[0010] A hybrid navigation system merges inertial location data from the inertial navigation unit and satellite location data from the satellite signal receiver. This navigation system integrates one or more Kalman filters arranged so that the hybrid navigation is recalibrated on the satellite location data.

[0011] However, in parallel with the development of satellite signal receivers, decoy devices have appeared to deceive these satellite signal receivers (this is referred to as "decoying" or "spoofing" the receivers). Such a device comprises an electronic processing unit connected to a radiofrequency signal transmitter to transmit fraudulent signals having the characteristics of satellite signals. More precisely, the electronic processing unit is arranged to develop, from an actual initial position of a satellite signal receiver, fraudulent signals which, when picked up by the satellite signal receiver, lead the satellite signal receiver to calculate an erroneous position.The actual initial position of the satellite signal receiver can be detected, for example, by means of a laser rangefinder or communicated by the vehicle carrying the satellite signal receiver via the emission of location signals imposed by certain navigation rules, in particular air and maritime (ADS-B or AIS signals emitted by vehicles to communicate their position to their neighbors). For fraudulent signals to be taken into account by the satellite signal receiver, it is not sufficient to emit the fraudulent signals with a power greater than the original satellite signals. It is also necessary that the fraudulent signals have the same code phase and a Doppler effect in the same range as those of the satellite signals previously received by the satellite signal receiver.If the first fraudulent signal received is consistent with the position last calculated by the satellite signal receiver and with the satellite signals received previously, and if the subsequently received fraudulent signals are consistent with each other, the fraudulent signals will be used by the satellite signal receiver as if they were real satellite signals and the error on the real position of the receiver. satellite signals cannot be detected.

[0012] The Kalman filter of hybrid navigation systems is protected by an innovation test to detect aberrant measurements and reject them. However, if the fraudulent signals have sufficient coherence, then they can satisfy this innovation test and it is therefore possible to bring about a recalibration of the hybrid navigation on the decoy positions. Thus, the satellite positioning data which normally make it possible to compensate for the errors of the inertial positioning data over the long term are affected by the fraudulent signals becoming a source of navigation error.

[0013] SUBJECT OF THE INVENTION

[0014] The invention aims in particular to detect a decoying operation and, once the decoying operation has been detected, to provide non-decoyed navigation. Summary of the invention

[0015] To this end, according to the invention, a navigation method is provided, comprising the continuous calculation of an inertial location and, at successive current times, the obtaining of a satellite location to calculate a hybridized location by hybridization of the inertial location and the satellite location, characterized in that the method comprises: - the calculation of at least one instantaneous difference between the inertial location and the hybridized location at each current instant; - the calculation, from each instantaneous deviation, of a deviation maintained from the instant of calculation of said instantaneous deviation until the current instant by an integration of a statistical error indicator of the location using a location error evolution model, and the storage of the maintained deviation to have, for a given period of time, a plurality of maintained deviations and statistical error indicators maintained from different calculation instants until the same current instant; - the calculation for each maintained deviation of at least one detection indicator comprising a ratio between: • a difference between the instantaneous gap and the sustained gap, and • a value based on the maintained statistical error indicator and a statistical error indicator of the hybridized location at the current time; - comparing, with a threshold, said detection indicator to detect a failure of the hybridized localization at the current time.

[0016] Thus, the deception is detected by verifying that the difference between, on the one hand, the instantaneous difference between the operational hybrid navigation and the inertial navigation at the current time and, on the other hand, each of the maintained differences is consistent with the errors specific to inertial navigation. Since several deviations are available, maintained from different times, a temporal analysis of the deviations can be carried out, which makes it possible to detect both fast decoy operations and slow decoy operations. It should also be noted that, with the invention, it is not necessary to calculate a plurality of inertial navigations from different times, but to maintain the deviations of these navigations compared to hybrid navigation, which requires simpler calculations that are significantly less costly in terms of resources.

[0017] Other characteristics and advantages of the invention will emerge from reading the following description of a particular and non-limiting embodiment of the invention. Brief description of the drawings

[0018] Reference will be made to the accompanying drawings, among which:

[0019] [Fig-1] [Fig.l] is a partial schematic view of an aircraft equipped with a navigation device implementing the method of the invention;

[0020] [Fig.2] [Fig.2] is a flowchart illustrating the method of the invention;

[0021] [Fig.3] [Fig.3] is a timing diagram illustrating the state of the memory of this device navigation at a given time. DETAILED DESCRIPTION OF THE INVENTION

[0022] With reference to the figures, the invention is described here in an aeronautical application, in relation to a navigation device on board an aircraft 1.

[0023] The navigation device comprises an inertial measurement unit 10, a satellite positioning signal receiving unit 20 and a baroaltimetric measurement unit 30, all three connected to an electronic navigation calculation unit 40.

[0024] The inertial measurement unit 10 comprises inertial sensors, namely linear inertial sensors (more precisely accelerometers 11.1, 11.2, 11.3) arranged along the x, y, z axes of a measurement frame m to measure the “gravitational speed” of this frame (i.e. the time integral of the specific force present at the center of this frame) and angular inertial sensors, here gyrometers 12.1, 12.2, 12.3 arranged along the x, y, z axes of this frame to measure the rotation of the measurement frame m relative to an inertial frame i not shown. The inertial sensors do not provide absolute values ​​but increments representative of a variation of the measured quantity relative to the previous measurement. The inertial reference frame is, for example, the measurement reference frame m when the inertial measurement unit 10 is switched on or any other inertial reference frame angularly offset from the latter.The increments of the specific force integral (denoted Incr.fs in [Fig.2]) are representative of a variation of the components of the gravitational velocity along the three axes of the measurement frame. The rotation increments (denoted Incr.co in [Fig.2]) are representative of the variation of the integral over time of . the angular rotation speed of the measurement frame m relative to the inertial frame and are provided in the form of quaternions, Euler angles, rotation matrices, or Bortz vectors. The inertial measurement unit 10 thus provides the electronic processing unit 40 with inertial positioning signals containing first data representative of a variation in gravitational speed (accelerometric measurement) and second data representative of a variation in angle (gyrometric measurement). Such an inertial measurement unit is known in itself and will not be detailed further here.

[0025] The satellite positioning signal reception unit 20 comprises an electronic circuit and an antenna for receiving and decoding satellite positioning signals originating from satellites belonging to a constellation of satellites in orbit around the Earth such as the constellations of the GPS, GALILEO, GLONASS, BEIDOU satellite positioning systems, etc. The electronic circuit is arranged to calculate pseudo-distances (PseudoD in [Fig. 2]) separating the satellite positioning signal reception unit 20 from each satellite from which a signal has been received at a time t and to provide the electronic processing unit 40 with a satellite location which here comprises a set of pseudo-distances at time t.Alternatively, the satellite positioning signal receiving unit 20 may be arranged to calculate a satellite position of the aircraft 1 from the pseudo-distances and to provide this satellite position as a satellite location. Such a satellite positioning signal receiving unit is known per se and will not be described in more detail here.

[0026] The baro-altimetric unit 30 is arranged to provide a signal AB (see [Fig.2]) representative of the altitude of the aircraft 1 as a function of the atmospheric pressure surrounding it. Such a baro-altimetric unit is known in itself and will not be detailed further here.

[0027] The electronic processing unit 40 comprises an electronic circuit having inputs connected to the outputs of the inertial measurement unit 10, of the satellite positioning signal reception unit 20 and of the baro-altimetric unit 30 for example by electrical conductors. The electronic processing circuit here comprises at least one processor and a memory containing a computer program which is executable by the processor and which comprises instructions arranged to implement the method of the invention.

[0028] More precisely, the electronic processing unit 40 implements a pure inertial navigation algorithm (IP in [Fig.2]) arranged to exploit the signals coming from the inertial measurement unit 10 to calculate an inertial location Loc.i of the aircraft 1, said location comprising: - a position (latitude, longitude and altitude), - a speed (or the vertical speed and two horizontal rotation components around the Earth), - an attitude (roll, pitch, yaw angles).

[0029] The baro-altimetric information AB is used to perform altitude feedback so as to have a more precisely determined altitude and a non-divergent vertical speed affected by a limited error, and to take into account the apparent curvature of the Earth's ellipse which depends on the altitude.

[0030] The electronic processing unit 40 also implements a hybrid navigation algorithm (Hyb in [Fig.2]) to calculate, in a manner known per se, a hybrid location Loc.h from the pseudo-distances PseudoD and the inertial location Loc.i. The hybridization consists, roughly, of resetting the pure inertial navigation to a satellite position of the aircraft determined from the pseudo-distances PseudoD. The hybridization methods are known per se and for example implement Kalman filters: they will not be detailed here.

[0031] The method according to the invention thus comprises, in a conventional manner, the continuous calculation of an inertial location and at a current instant, the obtaining of satellite location data to calculate a hybridized location by hybridization of the inertial location and the satellite location. The hybridized location, normally more precise than the pure inertial location, is used for the operational navigation of the vehicle outside of any decoy operation. It is understood that the current instant has been preceded in time by other current instants which have become past instants.

[0032] But the method of the invention also has the objectives of: - determining whether a decoying operation of the satellite positioning signal reception unit 20 is in progress, which is likely to affect the actual accuracy of the hybridized inertial location compared to the self-estimated accuracy of said hybridized inertial location, - if necessary, provide backup navigation not affected by the decoy.

[0033] We are interested here in horizontal decoy which is the most difficult to detect (the vertical error induced by a vertical decoy can more easily be detected from the barometric altitude AB). The location data which interest us are therefore the horizontal location data, namely the position data along the two horizontal axes and the speed data along these same two axes.

[0034] The method comprises the steps of: - calculate at least one instantaneous difference ELH_inst between the pure inertial horizontal location at the current time and the hybridized horizontal location at the current time; - then calculate, from the instantaneous difference ELH_inst, a difference ELH_entr maintained from the instant of calculation of said instantaneous deviation until the current instant by an integration of a statistical error indicator of the inertial localization, and storing in a memory (denoted MEM in [Fig.2]) the maintained deviation ELH_entr to have, for a given period of time, a plurality of maintained deviations and maintained statistical indicators, until the current instant, from different calculation instants (more precisely, from current calculation initialization instants which have since become past calculation initialization instants); - calculate, from these deviations and the statistical error indicators, a detection indicator Id which will be compared to a threshold S to detect a failure of the hybridized localization; - provide, in the event of the threshold being exceeded, emergency navigation which allows the hybrid navigation to be reset more precisely than pure inertial localization.

[0035] We therefore calculate deviations between the pure inertial navigation maintained for a time tr and the pure inertial navigation initialized by the hybrid navigation called “coasting” inertia since tr. For each of these deviations, the statistical error indicator is here the error covariance of the current hybridized horizontal location and the error covariance of the inertial horizontal location in coasting since tr. This covariance is initialized according to the covariance of the hybridized location error at the initialization time h and is then maintained via a linearized error model in the form of a horizontal location error evolution matrix and possibly UMI error parameters and random noise level of the evolution model.

[0036] It is understood that with the method of the invention, there are a plurality of differences between the pure inertial navigation initiated and the hybrid navigation from different predetermined times (and therefore from different satellite locations).

[0037] We note: - C (i varying from 1 to N+l, tl being the oldest instant and tN+i being the current instant) the date, in navigation time base, of the sampling instant of the horizontal location data used for the calculation and maintenance of the deviations; - At the time step, equal to or an integer multiple of the recalibration rate of the hybrid navigation filter, here equal to four seconds; - ELH(ti, tN+i) the maintained gap, from h to tN+[, between the inertial horizontal location and the hybridized horizontal location; - ELH(tN+i, tN+i) the instantaneous difference between the inertial horizontal location and the hybridized horizontal location at tN+i.

[0038] We understand that the hybridized horizontal location at the current time tN+[ a benefited from successive recalibrations on the satellite locations while the inertial horizontal location initiated from time h did not benefit from them.

[0039] In the absence of deception of the hybridized navigation, and neglecting the effect of the errors of the inertial measurement unit at the input of the pure inertial navigation, the evolution of the error of the pure inertial navigation in the absence of maneuver is perfectly deterministic at tN+i knowing: at ti, the exact values ​​of horizontal position error, the exact values ​​of projection error of the Earth's rotation (analytical platform position p relative to the Earth's reference frame), the exact values ​​of horizontal velocity around the Earth and the exact values ​​of attitude (analytical platform position p relative to the measurement reference frame m of the UMI's inertial sensors); the approximate values ​​of current latitudes and altitudes between h and tN+b

[0040] We therefore seek to verify that the difference between the instantaneous deviation ELH(tN+i, tN+i) and the maintained deviation ELH(t;, tN+i) at the current instant tN+[ is consistent with what is expected.

[0041] To obtain the deviation ELH(t;, tN+i), an integration is carried out, from the predetermined instant f to the current instant tN+b of the inertial localization variance using an inertial horizontal localization error evolution model. The inertial horizontal localization error evolution model is known in itself. It is generally a matrix F comprising seven error states: two error states on the angular position, two error states on the angular velocity of the movement of displacement around the Earth (or two error states of horizontal velocity), three error states of attitude.

[0042] More precisely here, the error evolution model used is the exponential of F*At, namely a linearized matrix of the horizontal inertial navigation error model at the average position and speed over the time step At. Taking into account the average position and duration over At in F makes it possible to improve the precision of the deviation calculation algorithm.

[0043] This model can be completed, in an approximate manner, to improve the accuracy of the error evolution model during acceleration, deceleration, or turning phases, by taking into account the effect of the azimuth error in the following manner: the acceleration input in the stabilized frame of the analytical platform p, constant is approximated by the variation of inertial horizontal speed (on At) on the speed error; we use a correction factor of the above approximation (ratio of the constant acceleration on At on the horizontal position error) equal to product of the azimuth error and the ratio of the deviation between the horizontal position variation (on At) and the horizontal position error.

[0044] More simply, we can calculate a propagation of the seven states and a desensitization noise (random walk of speed / attitude / position).

[0045] Models that can be used to implement the invention are described in the document "A Generic Inertial Navigation System Model for Computer Simulation Studies", DAB Gogg and RT Janus, Technical Memorandum WRSL-TM-30 / 90, Department of Defence, Defence Science and Technology Organisation, September 1990.

[0046] We recall that the location data which interest us here are the position data along the two horizontal axes and the speed data along these same two axes: we therefore calculate a position deviation ELH(t;, tN+i) for each of the two axes and a speed deviation ELH(t;, tN+i) for each of the two axes.

[0047] A detection indicator Id is calculated for each deviation ELH(t;, tN+i). Although for reasons of simplicity, we only speak here of a detection indicator, there is a detection indicator for the position and a detection indicator for the speed which are calculated in the same way.

[0048] The detection indicator Id is a ratio. The numerator is equal to the square of the difference between the instantaneous deviation ELH(tN+i, tN+i) and the sustained deviation ELH(t;, tN+i). The denominator is the sum of: - the horizontal localization variance resulting from the propagation of the horizontal localization error evolution model of the hybridized horizontal localization error covariances and inertial horizontal localization at h; - the horizontal localization error variance hybridized to tN+b

[0049] Each detection indicator Id is then compared to a threshold S set according to the desired false detection probability. The threshold is here set at 3.5 which allows a false detection probability here lower than 103.

[0050] If the detection indicator Id is lower than the threshold, the deviations are consistent and there is therefore a priori no deception. Hybrid horizontal localization remains used for operational navigation.

[0051] Otherwise, it is considered that deception is in progress and an alert is issued.

[0052] Since several deviations maintained since different times are available in memory, it is possible to carry out an analysis over the entire given period, which makes it possible to detect rapid deceptions causing a sudden drift in the hybrid horizontal location simply from the most recent deviations and slow deceptions causing a very gradual drift in the operational hybrid horizontal location from the oldest deviations.

[0053] It is therefore understood that the detection of slow decoys is all the more effective when old data is available. However, it is necessary to limit the number N of data in memory to limit its size and the volume of calculations to be carried out.

[0054] If we use a simple sliding time window to manage the contents of the memory (at each predetermined moment, the oldest data is erased and the last calculated data is stored), we understand that the time depth will be very limited.

[0055] We therefore preferentially use an algorithm for selecting the stored data such that after a certain time, the group of stored data contains, for the given period, a first series of maintained deviations recently stored and series of maintained deviations separated from the first series by multiples of the time step.

[0056] More precisely, the algorithm aims to keep in the group: - a first series of n maintained deviations, recently memorized, separated two by two by a time step, - a second series of n maintained deviations, separated from the first series by twice the time step, the maintained deviations of the second series being separated two by two by two time steps, - a third series of n maintained deviations, separated from the second series by four times the time step, the maintained deviations of the third series being separated two by two by four time steps, - a fourth series of n maintained deviations, separated from the third series by eight times the time step, the maintained deviations of the fourth series being separated two by two by eight time steps, - a fifth series of n maintained deviations, separated from the fourth series by sixteen times the time step, the maintained deviations of the fifth series being separated two by two by sixteen time steps, - a sixth series of n maintained gaps, separated from the fifth series by thirty-two times the time step, the maintained gaps of the sixth series being separated two by two by thirty-two time steps.

[0057] The number n is here equal to four. The first three series are shown in [Fig.3]: - the first series SI includes the deviations maintained since times -4, -8, -12, -16 (i.e. the current time minus one time step, the current time minus two time steps, the current time minus three time steps, the current time minus four time steps); - the gap maintained since the predetermined time -20 has been removed; - the second series S2 includes the deviations maintained since times -24, -32, -40, -48; - the gaps maintained since times -52, -56, -60, -64 have been removed; - the third series S3 includes the data of the predetermined instants -68, - 84,-100, 116; - and so on for the other series...

[0058] We can also present things like this, in this example with the number n equal to 4: - a value tj is selected every four values ​​to be propagated for tN+i-ti beyond 4*At (i.e. 16s if At is 4s); - on these values ​​propagated for tN+it; beyond 4*At, a value is selected every four values ​​to be propagated for tN+it; beyond 16*At (i.e. 64s if At is 4s); - on these values ​​propagated for tN+it; beyond 16*At, a value is selected every four values ​​to be propagated for tN+it; beyond 64*At (i.e. 256s if At is 4s); - on these values ​​propagated for tN+it; beyond 64*At, a value is selected every four values ​​to be propagated for tN+it; beyond 256*At (i.e. 1024s if At is 4s); - etc. - and finally no value is propagated beyond a maximum value of t N+l"ti.

[0059] This memory management mode allows for a much greater time depth than a sliding window for the same amount of data. Thus, with around forty maintained gaps stored, we have a time depth of around one hour.

[0060] It will be noted that, in the present description, the times for calculating the deviations have been neglected.

[0061] In the event of deception, having detection indicators corresponding to different past times of the start of the ELH deviation calculation also makes it possible to determine a time of start of deception (the oldest time for which the indicator Id is greater than the threshold).

[0062] It is therefore possible to calculate a new operational navigation by correcting the pure inertial navigation at the current time with the most recent maintained deviation prior to the time of start of decoying and propose it to the pilot of the vehicle.

[0063] A way of carrying out the calculations used for implementing the invention is detailed below.

[0064] To maintain the gaps, sampling is carried out over time intervals (typically 30 seconds to 2 minutes) and propagation is carried out over these intervals. of time, of: - N' sets of error states ELH(f, tN) of the pure horizontal inertial localization (which is assumed to be correct), - N' associated error covariance matrices P(ELH(t;, tN)).

[0065] At each sampling time step Ôt=tN+, tN, the following are carried out: - the calculation, once, of the matrix (loc.i(tN), loc.i(tN+i), ôt), exponential on the time interval ôt of the average matrix from tN to tN+i F(t) of inertial horizontal localization error (the vertical path being stabilized by means of the barometric altitude) which can be estimated as a constant on the time interval excluding accelerations (i.e. the specific force from which apparent gravity is subtracted); - the calculation at tN+i of the value of the vector ELH(tN+i, tN+i) of errors of the pure inertial horizontal localization (looped with the barometric altitude) compared to the hybridized localization at the same instant tN+[ ; - Sampling at the same time the (encompassing) covariance Pi(tN+i) of the hybridized localization errors at this time; - the N'-l calculations of the evolution of the error states Xj on the time step ôt using the formula Xj(tN+i)= <bXj(tN)+ <e>B*AV in which AV is the average horizontal velocity variation on ôt in the p frame of the inertial analysis platform. If the states defined in ELH are velocity errors, B is the 7x2 matrix which is 0 for all its components except for the two components corresponding to the two rows of the velocity error states in p and to the azimuth error column which are respectively + 1 / ôt and - 1 / ôt; - the N'-l calculations of evolutions of the error covariance matrices P(ELH(f, tN )) associated with the different values ​​of h available on the time step ôt using the formula

[0066] P(ELH(ti,tN+1))=P(ELH(ti,tN))*a>*P(ELH(ti,tN))'+C+Q*ôt

[0067] in which C is a corrective term corresponding to the azimuth error of the inertial localization in the presence of an acceleration and Q corresponds to the sum of the pseudo-random evolution of the errors and of a random walk in angle and speed of the inertial measurement unit (this formula is known in itself). The determination of the matrix B is a little different if the states defined in ELH are errors of rotation speed around the terrestrial ellipsoid in p, it is necessary to take into account the local curvature matrix of the ellipsoid in p (at most precisely its median value between tN and tN+p

[0068] In a generic manner, we can calculate for each time step an evolution matrix <b’ approximée précisément en fonction de la localisation (pas seulement horizontale) médiane et de la différence de localisation entre tN et tN+p

[0069] We set the formulas for calculating the evolution of tN to tN+i: ELH(ti,tN+i)))= '*ELH(ti,tN ))et

[0070] P(ELH(ti,tN+1))=P(ELH(ti,tN))* <D’*P(ELH(ti,tN))+Q*(tN+i-tN)

[0071] The calculation of the evolution of the horizontal location deviation covariance can be refined by supplementing the horizontal location error states with IMU error states (for example, residual bias and misalignments of the sensitive axes of the IMU inertial sensors) and by supplementing the evolution matrix (of dimension > 7x7) accordingly.

[0072] In the case of very large potential altitude differences between pure inertial navigation and hybrid navigation, another error term can also be taken into account for the calculation of the evolution of tN to tN+the covariances P(ELH(t;, tN)) due to the curvature difference as a function of altitude. An altitude difference of 320m generates, for a given horizontal speed, a horizontal position evolution difference of approximately 0.0001 times this speed (2*320 / RT ❖0.0001 with the mean radius of the Earth RT ❖ 6,400,000 m). This therefore represents a potential position error evolution of 3cm / s for a device at 300m / s, which is generally low or equivalent to the effect of the speed errors of hybrid navigation and the modeled errors of the UMI.

[0073] The use of location variation data between tN and tN+[ (essentially horizontal velocity) in addition to the median location allows for a more accurate calculation of location error variations from tN to tN+i.

[0074] The attitude values ​​at tN and tN+i can be used for the calculation of the evolution of the covariances from tN to tN+i of the error of the ELH(t;, tN) if errors of the inertial measurement unit (for example accelerometric or gyrometric biases or misalignments of the axes of the sensors) are modeled in addition to the inertial localization errors "ELH" to obtain a more precise evaluation of the error domains determined by the covariance matrices "P" as a function of time.

[0075] It is understood that at each “new ti”, sampling dates of the pure inertial localization and of the hybridized localization, we propagate from tN to tN+i all the localization deviation states ELH(t;, tN) maintained until tN and already available to obtain the values ​​ELH(ti, tN+i). We also recover the new set of localization deviation states ELH(tN+i, tN+i) i.e. the deviation states between the pure inertial localization and the hybridized localization at tN+i, and we possibly eliminate the surplus ELH(t;, tN+i) state(s).

[0076] The values ​​of the N' detection indicators are obtained by the formula:

[0077] Id(ti)=(Loc.i(ti)-H*ELH(ti,tN+1)-Loc.h(ti))2 / H*P(ELH(ti,tN+1))*Ht for all available f and ELH(ti, tN+i). P(ELH(t;, tN+i)) is the error matrix or error covariance of ELH(ti, tN+i). H corresponds to the observations, namely a horizontal position error and, in the preferred embodiment described here, a horizontal velocity error. rizontal. We note that tN+it; is the detection delay of Id(L).

[0078] Of course, the invention is not limited to the embodiment described but encompasses any variant falling within the scope of the invention as defined by the claims.

[0079] In particular, the time step used for storage is here that of the recalibrations but it can be different and for example correspond to a multiple of the time step of the recalibrations.

[0080] Filtering of the stored data is optional or can be different: for example, a classic sliding time window can be used to renew the group of stored data (the oldest data is removed from the group each time new data is stored).

[0081] Although hybrid navigation is calculated from pseudo-distances, it is possible to calculate hybrid navigation from a pure satellite position obtained from pseudo-distances.

[0082] Alternatively, the variance of the hybrid horizontal location can be used as a statistical error indicator. Indeed, as long as satellite signals are received, the variance of the hybrid horizontal location is negligible compared to the variance of the inertial horizontal location. On the other hand, the use of covariance improves the performance of the method of the invention when the satellite data sets are unavailable for at least a period of the order of a few minutes.

[0083] The invention can be implemented by using as horizontal location data only the position data along the two horizontal axes. However, it should be noted that taking into account the speed data along these same two axes makes it possible to improve the accuracy.

[0084] Locations can be provided in polar or Cartesian coordinates. < / e>

Claims

Claims

1.

2. Navigation method, including the continuous calculation of a location inertial location and, at successive current times, obtaining a satellite location to calculate a hybrid location by hybridization of the inertial location and the satellite location, characterized in that the method comprises: the calculation of at least one instantaneous difference between the inertial location and the hybridized location at each current instant; the calculation, from each instantaneous difference, of a difference maintained from the instant of calculation of said instantaneous difference until the current instant by an integration of a statistical error indicator of the location using a location error evolution model, and the storage of the maintained difference to have, for a given period of time, a plurality of maintained differences and statistical error indicators maintained from different calculation instants until the same current instant; the calculation for each maintained deviation of at least one detection indicator comprising a ratio between: a difference between the instantaneous deviation and the maintained deviation, and a value based on the maintained statistical error indicator and a statistical error indicator of the hybridized location at the current time; comparing, with a threshold, said detection indicator to detect a failure of the hybridized localization at the current time. The method of claim 1, wherein the error indicator sta tistic includes the variance of the inertial localization and in which the detection indicator is equal to the square of the difference between the instantaneous deviation and the maintained deviation on the sum of the variance of the maintained inertial localization and the variance of the hybrid localization at the current instant.

3. Method according to claim 1 or 2, comprising, in the event of exceeding the threshold, the provision of a backup navigation in which the hybridized location is replaced by a corrected inertial location with a deviation maintained since an instant prior to the detected failure.

4. A method according to any preceding claim, wherein the inertial localization uses horizontal localization data comprising position data.

5. The method of claim 4, the horizontal location data also includes speed data.

6. A method according to any preceding claim, wherein the satellite location uses satellite location data comprising pseudo-ranges.

7. A method according to any preceding claim, wherein the locations are calculated at stabilized altitude and vertical speed.

8. Method according to any one of the preceding claims, in which the storage of the deviations is carried out according to a predetermined time step.

9. Method according to claim 8, in which the deviations are filtered over time to keep in the group, for the given period, a first series of recently stored maintained deviations and series of maintained deviations of the first series by multiples of the time step.

10. Method according to claim 9, in which the stored deviations are filtered over time to keep in the group, for the given time period: - a first series of n maintained deviations last stored separated two by two by a time step, - a second series of n maintained deviations, separated from the first series by twice the time step, the maintained deviations of the second series being separated two by two by two time steps, - a third series of n maintained deviations, separated from the first series by four times the time step, the maintained deviations of the third series being separated two by two by four time steps, a fourth series of n maintained deviations, separated from the first series by eight times the time step, the maintained deviations of the fourth series being separated two by two by eight time steps, a fifth series of n maintained gaps, separated from the first series by sixteen times the time step, the maintained gaps of the fifth series being separated two by two by sixteen time steps, a sixth series of n maintained gaps, separated from the first series by thirty-two times the time step, the maintained gaps of the sixth series being separated two by two by thirty-two time steps.

Images