Three-dimensional positioning of an aircraft

Abstract

The present invention relates to a method (28) for three-dimensional positioning of an aircraft, which comprises the following steps: - obtaining (30) elementary radar measurements acquired at distinct measurement times of a predetermined acquisition period, and in a predetermined radar reference frame; - obtaining (32) inertial measurements acquired over the predetermined acquisition period; - synchronising (34) the elementary radar measurements and the inertial measurements by estimating the inertial measurements corresponding to the distinct measurement times which are representative of each of the elementary radar measurements; - projecting (36) each of the elementary radar measurements into a single unified reference frame associated with a single predetermined time of the predetermined acquisition period, using, for each of the elementary radar measurements, the estimated inertial measurement associated with the same acquisition time as the elementary radar measurement considered; and - three-dimensional positioning (38) of the aircraft by incoherent integration of the projection.

Description

[0001] TITLE: Three-dimensional localization of an aircraft

[0002] The present invention relates to a method for three-dimensional localization of an aircraft, said aircraft being equipped with at least one radar and at least one inertial measuring instrument.

[0003] The invention also relates to a computer program comprising software instructions which, when implemented by a programmable electronic device, implement such a process.

[0004] The invention also relates to a three-dimensional localization device for an aircraft, said aircraft being equipped with at least one continuous wave radar and at least one inertial measuring instrument.

[0005] The invention also relates to a landing aid system comprising at least one radar and such a device.

[0006] The invention also relates to an aircraft comprising such a landing aid system.

[0007] The invention relates to the field of aircraft navigation. In this field in particular, it is known to equip aircraft with landing aid devices capable, in particular, of determining the aircraft's position relative to the landing runway.

[0008] In this field, the "post-integration" (also known as "non-coherent integration" or "incoherent integration") of elementary measurements (i.e., echoes) of a radar, particularly a pulse radar, over a fixed time interval makes it possible to improve the estimation of contrasts and the localization of points of interest (POI).

[0009] By "post-integrations" we mean the quadratic sums of the phase I and quadrature Q radar measurements after the coherent sums of phase I and quadrature Q.

[0010] Whether the pilot is human or automated, the performance required to provide localization assistance functions during approach for landing requires a post-integration period on the order of several tenths of a second, or even several seconds.

[0011] In the case of application to the localization of an aircraft, particularly during approach and landing phases, the effectiveness of post-integration is not optimal with regard to the normative aeronautical requirements for approach and landing assistance, as radar measurements, particularly continuous wave radar, are too spread out over time, and / or too altered in accuracy by the movements of the aircraft to be located, the reference point for elementary measurements varying according to the movements and rotations of the aircraft.

[0012] In other words, currently determining the attitude changes of an aircraft autonomously by radar, without inertial assistance or GNSS (Global Navigation Satellite System), is hampered by the accuracy of elementary measurements, so that, for this application case, radar measurements are degraded by the movements of the aircraft to be located and / or detected.

[0013] The aim of the invention is therefore to propose a more efficient localization solution with post-integration (i.e. incoherent integration), allowing the dynamics of the aircraft to be taken into account.

[0014] To this end, the invention relates to a method for three-dimensional localization of an aircraft, said aircraft being equipped with at least one radar and at least one inertial measuring instrument, said method comprising at least the following steps:

[0015] - obtaining, via said at least one radar, elementary radar measurements acquired at measurement times distinct from a predetermined acquisition period, each measurement being acquired in a predetermined radar reference frame associated with said radar and associated with said instant of its acquisition;

[0016] - obtaining, via said at least one inertial measuring instrument, inertial measurements acquired over said predetermined acquisition period;

[0017] - synchronization of said elementary radar measurements and said inertial measurements by estimation of the inertial measurements corresponding to the distinct measurement instants representative of each of said elementary radar measurements;

[0018] - projection of each of said elementary radar measurements into the same unified frame associated with the same predetermined instant of said predetermined acquisition period, using, for each of said elementary radar measurements, the estimated inertial measurement associated with the same acquisition instant as said elementary radar measurement considered;

[0019] - three-dimensional localization of said aircraft by inconsistent integration of said projection.

[0020] Thus, the present invention makes it possible to reduce, or even eliminate, the degradation of radar measurements due to the movements of the aircraft to be located and / or detected by harmonizing these radar measurements with inertial measurements.

[0021] To do this, the said inertial measurements are synchronized and / or estimated on the acquisition times of the radar measurements, then the radar measurements are brought back into the same frame associated with the same predetermined instant of the acquisition period (i.e. the post-integration period) of the said radar measurements by using, for each of the said elementary radar measurements, the estimated inertial measurement associated with the same acquisition instant as the said elementary radar measurement considered.

[0022] The use of the inertial measurement respectively associated with each of the said elementary radar measurements makes it possible to take into account the rotational and translational movements of the aircraft during the post-integration acquisition period (i.e. non-coherent integration).

[0023] Harmonizing radar measurements with inertial measurements improves the efficiency of the localization solution with post-integration (i.e., inconsistent integration), particularly when used in aeronautics to descend below 200 feet above ground level and land.

[0024] Indeed, harmonizing radar measurements using inertial measurements increases the amplitude and precision of radar measurement contrasts through post-integration.

[0025] According to other advantageous aspects of the invention, the three-dimensional localization method for an aircraft comprises one or more of the following features, taken individually or in all technically possible combinations:

[0026] - said estimation includes a polynomial interpolation taking into account, for an elementary radar measurement acquired at an acquisition time, the first inertial measurement acquired after said acquisition time and a predetermined number of previous inertial measurements;

[0027] - said predetermined instant of said predetermined acquisition period is the acquisition instant of the last elementary radar measurement acquired on said predetermined acquisition period;

[0028] - said projection includes at least one weighting determined according to the predetermined quantification of the measurement axes of said at least one radar.

[0029] - said process includes a step of recalibrating said unified reference frame by tracking at least one point of interest during said predetermined acquisition period and the preceding predetermined acquisition period;

[0030] - said recalibration is performed using a weighted least squares function or by filtering of a type belonging to the group comprising at least:

[0031] - a Kalman filter;

[0032] - Bayesian filtering;

[0033] - particle filtration.

[0034] The invention also relates to a computer program comprising software instructions which, when implemented by a programmable electronic device, implement such a method. The invention also relates to a three-dimensional aircraft localization device, said aircraft being equipped with at least one continuous wave radar and at least one inertial measurement instrument, said device comprising:

[0035] - a first acquisition module configured to obtain, via said at least one radar, elementary radar measurements acquired at measurement times distinct from a predetermined acquisition period, each measurement being acquired in a predetermined radar reference frame associated with said radar and associated with said instant of its acquisition;

[0036] - a second acquisition module configured to obtain, via said at least one inertial measuring instrument, inertial measurements acquired over said predetermined acquisition period;

[0037] - a synchronization module configured to synchronize said elementary radar measurements and said inertial measurements by estimating the inertial measurements corresponding to the distinct measurement instants representative of each of said elementary radar measurements;

[0038] - a projection module configured to project each of said elementary radar measurements into the same unified frame associated with the same predetermined instant of said predetermined acquisition period, using, for each of said elementary radar measurements, the estimated inertial measurement associated with the same acquisition instant as said elementary radar measurement considered;

[0039] - a three-dimensional localization module configured to locate said aircraft by inconsistent integration of the projection provided by the projection module.

[0040] The invention also relates to a landing aid system comprising at least one radar and such a device.

[0041] The invention also relates to an aircraft comprising such a landing aid system.

[0042] The invention will become clearer upon reading the following description, given solely by way of non-limiting example, and made with reference to the drawings in which:

[0043] [Fig. 1] Figure 1 is a block diagram of the main functional blocks of an electronic three-dimensional aircraft localization device according to the present invention;

[0044] [Fig 2] Figure 2 is a flowchart of the main steps of a three-dimensional localization method for an aircraft according to the present invention;

[0045] [Fig 3] Figure 3 illustrates the synchronization implemented according to the present invention;

[0046] [Fig 4] Figure 4 illustrates reference points used according to the present invention. In the following description, the expression "approximately equal to" is understood as a relationship of equality to plus or minus 10%, that is to say with a variation of at most 10%, preferably also as a relationship of equality to plus or minus 5%, that is to say with a variation of at most 5%.

[0047] Figure 1 illustrates an embodiment of an electronic three-dimensional aircraft localization device according to an embodiment of the invention.

[0048] Such a device is suitable for integration within a landing aid system, not shown, comprising at least one radar, in particular continuous wave or even pulsed, not shown, said landing system itself being suitable for being carried on board an aircraft not shown.

[0049] More specifically, the electronic three-dimensional localization device 10, according to the present invention, is suitable for implementing the three-dimensional localization method of an aircraft which will be described later in relation to Figure 2.

[0050] To do this, such a device 10 first includes a first acquisition module 12 configured to obtain, via said at least one radar, elementary radar measurements acquired at measurement times distinct from a predetermined acquisition period, each measurement being acquired in a predetermined radar frame associated with said radar and associated with said instant of its acquisition.

[0051] Furthermore, the device 10 includes a second acquisition module 14 configured to obtain, via said at least one inertial measuring instrument, inertial measurements acquired over said predetermined acquisition period.

[0052] Next, the device 10 includes a synchronization module 16 configured to synchronize said elementary radar measurements and said inertial measurements by estimating the inertial measurements corresponding to the distinct measurement instants representative of each of said elementary radar measurements.

[0053] In addition, the device 10 includes a projection module 18 configured to project each of said elementary radar measurements into the same unified frame associated with the same predetermined instant of said predetermined acquisition period, using, for each of said elementary radar measurements, the estimated inertial measurement associated with the same acquisition instant as said elementary radar measurement considered.

[0054] Finally, the device 10 includes a three-dimensional localization module 20 configured to locate said aircraft by inconsistent integration of the projection provided by the projection module.

[0055] In the example shown in Figure 1, the electronic three-dimensional positioning device 10 of an aircraft comprises a processing unit 22 consisting, for example, of a memory 24 and a processor 26 associated with the memory 24. The device 10 is at least partially implemented as software, or a software component, executable by the processor, including the first acquisition module 12, the second acquisition module 14, the synchronization module 16, the projection module 18, and the three-dimensional positioning module 20. The memory 24 of the electronic three-dimensional positioning device 10 of an aircraft is thus capable of storing such software or software components, and the processor 26 is thus capable of executing them.

[0056] In an unrepresented variant, the first acquisition module 12, the second acquisition module 14, the synchronization module 16, the projection module 18 and the three-dimensional localization module 20 are each implemented as a programmable logic component, such as an FPGA (Field Programmable Gate Array), or as a dedicated integrated circuit, such as an ASIC (Application Specific integrated Circuit).

[0057] When a portion of the electronic three-dimensional localization device 10 for an aircraft within a landing scene according to the present invention is implemented in the form of one or more software programs, i.e., in the form of a computer program, this portion is also capable of being stored on a computer-readable medium (not shown). The computer-readable medium is, for example, a medium capable of storing electronic instructions and being connected to a bus of a computer system. For example, the readable medium is an optical disc, a magneto-optical disc, a ROM, a RAM, any type of non-volatile memory (e.g., EPROM, EEPROM, FLASH, NVRAM), a magnetic card, or an optical card. A computer program containing software instructions is then stored on the readable medium.

[0058] An example of the operation of the electronic three-dimensional localization device 10 of an aircraft according to the present invention, as previously illustrated in relation to Figure 1, is described below in relation to Figure 2.

[0059] More specifically, the method 28 for three-dimensional localization of an aircraft generally includes a first step 30 of obtaining O_MR, via said at least one radar, elementary radar measurements acquired at measurement times distinct from a predetermined acquisition period, each measurement being acquired in a predetermined radar frame associated with said radar and associated with said instant of its acquisition.

[0060] In parallel, or prior to, or successively, the method 28 for the three-dimensional localization of an aircraft includes a second step 32 of obtaining O_MI, via said at least one inertial measuring instrument, inertial measurements acquired over said predetermined acquisition period. Once the two preceding steps 30 and 32 have been carried out, the method 28 for the three-dimensional localization of an aircraft includes a step 34 of synchronizing said elementary radar measurements and said inertial measurements by estimating the inertial measurements corresponding to the distinct measurement instants representative of each of said elementary radar measurements. This step will be described in more detail later with reference to Figure 3.

[0061] Then, the three-dimensional localization method 28 of an aircraft includes a step 36 of projecting each of said elementary radar measurements into the same unified frame associated with the same predetermined instant of said predetermined acquisition period, using, for each of said elementary radar measurements, the estimated inertial measurement associated with the same acquisition instant as said elementary radar measurement considered.

[0062] Finally, the method 28 for three-dimensional localization of an aircraft includes a step 38 for three-dimensional localization L of said aircraft by inconsistent integration of said projection.

[0063] As an optional addition, the said predetermined instant of the said predetermined acquisition period is the acquisition instant of the last elementary radar measurement acquired on the said predetermined acquisition period.

[0064] Subsequently, we describe in more detail certain variants implemented during the synchronization steps 34 and projection steps 36.

[0065] More specifically, as previously stated, said synchronization step 34 includes a step 40 of estimation of each inertial measurement respectively associated with the same acquisition instant as each elementary radar measurement considered.

[0066] According to an optional variant shown in dotted lines, said estimation 40 includes a polynomial interpolation substep 42 I which takes into account, for an elementary radar measurement acquired at an acquisition time, the first inertial measurement acquired after said acquisition time and a predetermined number of previous inertial measurements.

[0067] It should be noted that, in a specific case, it is possible to dispense with this optional interpolation step when at least one inertial measurement instrument is integrated into the radar and synchronized with the radar measurements (i.e., with a single time base for both radar and inertial measurement sub-functions). This simplifies the calculations described previously and eliminates the need for this synchronization step. In other words, in this specific case, the architectural constraint, due to the integration of the inertial measurement instrument within the radar, ensures the synchronization function. As an optional supplement represented by dashed lines, projection 36 includes at least one weighting step 44 determined according to the predetermined quantization of the measurement axes of at least one radar.

[0068] According to another optional supplement shown in dotted lines, the method 28 for three-dimensional localization of an aircraft includes a step 46 of re-registering R said unified reference frame by tracking at least one point of interest during said predetermined acquisition period and the preceding predetermined acquisition period.

[0069] According to the embodiment variant of Figure 2, the registration 46 is a sub-step of the projection step 36 (i.e. the registration is implemented during the projection step).

[0070] According to another variant not shown, the optional recalibration 46 is implemented prior to at least the said projection step 36.

[0071] As an optional addition, this recalibration is performed using a weighted least squares function or by filtering of a type belonging to the group comprising at least:

[0072] - a Kalman filter;

[0073] - Bayesian filtering;

[0074] - particle filtration.

[0075] Figure 3 illustrates more precisely, by means of example 50, the synchronization implemented according to the present invention.

[0076] Indeed, the principle of harmonizing radar measurements using inertial measurements proposed according to the present invention consists first of all in synchronizing inertial measurements on the representative instants of the elementary measurements of the radar.

[0077] Note that in the case of a continuous wave radar, these elementary measurements are the product of coherent integrations carried out over a short period (as such, they are sometimes called instantaneous).

[0078] Inertial measurements are obtained via said at least one inertial measurement instrument. Note that such an inertial measurement instrument corresponds, according to a first variant, to an INS or INU (from the English inertial navigation system or inertial measurement unit respectively), or according to a second variant to a GNSS (from the English Global Navigation Satellite System) receiver with interferometry, as notably described in 1991 by F. Van Graas et al. in the article entitled "GPS Interferometric Attitude and Heading Determination: Initial Flight Test Results".

[0079] The objective of synchronization step 34 in Figure 2, an example of which is illustrated in Figure 3, is to prepare for the harmonization via the subsequent step 36 of projecting radar measurements using inertial measurements to be as reliable as possible. In principle, synchronization step 34 in Figure 2, an example of which is illustrated in Figure 3, consists of identifying and establishing the relationships between the representative times of the radar measurements on the one hand and the inertial measurements on the other.

[0080] Three distinct time scales are shown in Figure 3, namely the 52 time scale associated with elementary radar measurements acquired in the radar frame, illustrated later in relation to Figure 4, at successive times of index n, with n varying between 1 and N.

[0081] Time scale 54 is associated with inertial measurements acquired in the "platform" frame, namely the frame of said at least one inertial measuring instrument.

[0082] Time scale 56 is associated with the capture of a radar image or measurement intended for a positioning aid function, considering that a radar image or measurement intended for a positioning aid function, for a human pilot or an autopilot, consists of a set of elementary radar measurements that are spread over a time period 8t k defined by the following localization time interval: [t k-1 = t k - 8t k ; t k ], as illustrated by time scale 56.

[0083] In other words, we obtain a radar image at a current instant t k-1 , then a subsequent image at a given time t k , the moments t k-1 and t kbeing separated by the time period 8t k .

[0084] During this period of time 8t k As illustrated by timescale 52, there are Q series of elementary radar measurements (i.e., Q series of pulses, i.e., Q bursts of pulses) associated with each radar image intended for a localization aid function.

[0085] In the example of Figure 3, we consider that Q = 4 series of three burst pulses with 8t n the period of a pulse (i.e., an elementary radar measurement) within a series of elementary radar measurements. In Figure 3, within the third series of elementary radar measurements (i.e., the third series of pulses), time t n is associated with the beginning of the third pulse, and this third series of elementary radar measurements is then centered on the instant represented by the corresponding vertical dashed line

[0086]

[0087] In other words, in temporal terms, each elementary measurement is representative of the midpoint of a period 8t n whose end time is noted t n Thus, the temporal discretization of radar measurements, relative to time t n -

[0088]

[0089] : (n)

[0090] As previously mentioned, according to one variant, elementary radar measurements (i.e., pulse series) are typically ordered into "bursts" (temporal structuring) and correspond, for example, to "FMCW chirps," namely pseudo-periodic signals frequency-modulated around a carrier frequency and also amplitude-modulated by an envelope whose variations are slow compared to the phase oscillations. Indeed, since it is generally difficult to transmit random frequencies "cleanly," frequency-modulated continuous wave (FMCW) radars gradually vary the frequency of their signal using ascending and descending ramps.

[0091] FMCW chirps, for example, are typically repeated sawtooth signals used to accumulate elementary measurements (increasing reflector contrast relative to radar clutter) and to scan all the circulars (and / or elevations) of the antenna's angular aperture. The circular of a predetermined point corresponds to the angle formed from the radar's line of sight to the projection of said predetermined point onto the plane of projection of the circulars. This plane is defined by the radar's line of sight and the axis of angular measurements of said radar reference.

[0092] The "bursts" are not necessarily continuous temporally as for example in the case of FMCW duty cycles, which is schematically represented on time scale 52 in figure 3 by the spaces separating the bursts of three pulses.

[0093] The St temporal stepn (i.e., the duration of a pulse) is adjusted according to the "instantaneous" criterion of the quantified elementary measurement with regard to the angular accuracy requirements, namely 10% of 0.3° at a rate of 10⁷s, which gives St n < 3 ms, in order to limit the real-time load (interpolations of synchronizations and changes of reference of harmonizations).

[0094] In this case, by extension, δt n no longer represents just a single "elementary" radar measurement, but a set of such measurements juxtaposed temporally. For example, and typically: δt n would then represent a whole series of elementary measurements, including three elementary measurements in Figure 3 and sixteen according to a more realistic example, instead of just one."

[0095] In other words, according to the example in Figure 3, in terms of synchronization, the time period St k inter-image is such that its starting instant t k-1coincides with the moment t n first - Stn, °ù t n irst is the start time of the first elementary radar measurement of index n=1 acquired after time t k-1 from the beginning of the St time period k inter-image (i.e., the first pulse of the first series of successive burst pulses at time t) k-± ), while the end time t k of the St time period k inter-image corresponds to time t n_last from the end of the last pulse of the last elementary radar measurement (i.e. the last pulse of the last burst pulse series associated with said image).

[0096] In other words, t n _ first = t n=1 and t n_last = t n=N ..

[0097] As illustrated by timescale 54, inertial, attitude and / or velocity measurements are, for example, acquired periodically with a time period of 8t sbetween each inertial measurement, with, in the example of Figure 3, t s first the instant of the first inertial measurement belonging to the time period 8t k inter-image, t s the instant of the sixth inertial measurement belonging to the time period 8t k inter-image, t s_last the instant of the last inertial measurement to be taken into account for harmonization, this inertial measurement being the first inertial measurement that is carried out after the expiration of the time period 8t k inter-image (i.e., the first inertial measurement that follows in time the end instant t) k of the time period 8t k inter-image).

[0098] Inertial measurements are preferably considered unfiltered, corrected for the effect called "coning", a parasitic effect induced by two orthogonal and non-hybridized rotations, therefore directly derived from inertial gyroscopic increments.

[0099] The synchronization implemented according to the present invention consists of estimating, according to step 40 of Figure 2, inertial measurements at times representative of the corresponding radar measurements, using the inertial measurements acquired between times t s_first and t s_last .

[0100] For example, we use the inertial measurement obtained at time t s to determine an estimate of the inertial measurement at the representative instant of the third elementary radar measurement shown in dashed lines t = t n -

[0101]

[0102] or, by approximation, in t n .

[0103] As previously mentioned in relation to Figure 2, according to one variant, the estimation of inertial measurements will preferably be obtained by interpolation (while waiting for the inertial measurement at time t). s following the elementary radar pulse acquired at t nSuch an interpolation is preferentially polynomial (i.e., taking into account, for example, the two to four previous inertial measurements).

[0104] Note that optionally, in the case where the inertial measurements are filtered upstream, then the pure delay due to the filter is likely to be compensated during the estimation step 40.

[0105] Thus, at the end of the synchronization step, we obtain, for example, the following inertial measurements, interpolated at time t n , and acquired in the "platform" frame of reference, namely the frame of reference of said at least one inertial measuring instrument:

[0106] - the aircraft's attitudes; with ip the defined yaw angle

[0107]

[0108] relative to the deviation from geographic North, called the wander angle (or relative to geographic North), positive along the rotation around the Z-axis, 6 the pitch angle, defined relative to the Earth's horizontal plane, positive in the nose-up position, and <p l’angle de roulis, défini par rapport au plan horizontal terrestre, positif si penché vers la droite;

[0109] - the speeds:

[0110]

[0111] with [p(n)] the platform frame of said at least one inertial measuring instrument at time t n .

[0112] Note that maintaining the verticality of the Z axis of this "platform" reference frame is ensured by the identification of gravity, or alternatively via GNSS attitude measurements, notably described by J. L Gerner in the document entitled "Topstar 300 - An Enhanced GPS receiver for Space Applications" of November 2000.

[0113] Inertial velocity measurements are typically derived from inertial accelerometer increments, refined or replaced by radar measurements or measurements derived from GNSS.

[0114] Figure 4 illustrates reference points used according to the present invention, particularly during projection step 36 of Figure 2.

[0115] More precisely, according to the example illustrated in Figure 4, a predetermined radar reference frame 60 is used, whose origin O corresponds to the phase center of the radar antenna. Furthermore, the x-axis of the predetermined radar reference frame 60 is aligned with the radar's line of sight and inclined at an angle y (not shown and called the radar tilt) with respect to the aircraft's lubber line; this tilt angle is defined as positive if the radar is pointing above the lubber line (i.e., a direct orthonormal reference frame).

[0116] The y-axis of the ordinates in the 62 projection plane of the circulars is perpendicular to this x-axis of the line of sight, and parallel to the line of the transmitters (i.e., the width of the radar), and in particular oriented to the right when looking along the radar line of sight. The z-axis of the altitudes, according to this example, is such that the radar coordinate system 60 is orthonormal right-handed (i.e., downwards if the aircraft's attitude is zero).

[0117] Thus, as illustrated by Figure 4, in such a radar frame 60, we consider M, a predetermined point, defined by its distance to the radar p and its angles a and p such that: a being the circular of M as defined previously and the elevation of said predetermined point corresponding to the opposite of the angle formed from the vector starting from the origin of the radar frame to said predetermined point M in the plane 64 (O, M, z), up to the projection of this vector in the plane 62 of projection of the circulars.

[0118] The Cartesian coordinates of M in such a radar frame 60, also called frame R, also referred to hereafter as [r], are then such that: x [r] = p • cos (a) • cos (?),

[0119]

[0120] More specifically, the radar provides elementary radar measurements in amplitude A and phase 0 relative to a set of geometric configurations, also called a "radar grid", fixed by considering:

[0121] - p the distance of the measurement point, discretized into "distance box (i)" by steps of the distance resolution in a definition domain corresponding to the radar range, and

[0122] - a circular (also called "bearing") of the measurement point, corresponding, as previously indicated, to the angle formed from the origin and the line of sight (0, x)[ r ] of the radar, by rotation in the (0, x, y) plane [R around the (0, z) axis r], and discretized into "circular box (j)" by step of the circular resolution, in a domain of definition following the angular opening of the circular radar, the plane of the circular being perpendicular to (0, x, y)[r] and including the direction pointed by a.

[0123] In complex exponential form, the elementary radar measurement Mes_Radar can then be denoted as follows (f mis for function):

[0124]

[0125] Optionally, the geometric configurations, also called "radar grid", defining the elementary radar measurements are suitable to be completed, by p the aforementioned elevation corresponding to the angle formed, in the plane of the circular a, by the measurement at the distance p from 0 and its projection in the plane (0, x, y) [r], this angle being discretized into "distance case (e)" by steps of the resolution in "elevations" in a domain of definition following the angular opening of the radar in elevations.

[0126] Note that if the elevation of the elementary radar measurements is not available, it can be determined from a measurement of the distance Z>(n)[r(n)] to the ground along the line of sight from the phase center of the radar [(0, x)[ r j] at time t n , according to the procedure described in application FR 3152598 on behalf of the Applicant.

[0127] Optionally, the geometric configurations, also called "radar grid", defining the elementary radar measurements are suitable for being completed, by the Doppler of the elementary radar measurement, discretized into "Doppler box (d)" by steps of the Doppler resolution in a domain of definition following the Doppler coverage of the radar measurements.

[0128] Note that, as an optional complement, an aircraft landing assistance system comprising at least one radar and an electronic three-dimensional localization device according to the present invention, is configured to implement said at least one radar and said device using a sliding implementation window.

[0129] In other words, the method according to the present invention is suitable for implementation using a sliding window (FIFO, First in First) for measurements and calculations, particularly for integrals of

[0130]

[0131] To ensure the required refresh rate (e.g., 5 to 15 Hz depending on the application), while maintaining a measurement period greater than this time, a sliding window will need to be applied.

[0132] Furthermore, with regard to basic radar measurements, it is necessary to maintain the values ​​and discretization steps throughout the performance of the functions to be performed (e.g. especially in the case of an approach for landing) and for this, the domain of definition of distances and circulars is extended according to the actual movements of the aircraft during the duration of the function to be performed.

[0133] Thus the distance and circular "boxes" of the first elementary measurements are designed to persist throughout the function to be performed, with, for example, a horizon of the order of one minute in the case of an approach for landing.

[0134] According to the present invention, an intermediate consideration is also given to a predetermined geographical reference frame [g] having an origin corresponding to that of the reference frame of said at least one inertial measuring instrument, an abscissa axis along the terrestrial geographic North direction, an ordinate axis along the terrestrial geographic East direction and an altitude axis along the downward-oriented local geographic terrestrial vertical.

[0135] The platform reference frame [p] of said at least one inertial measuring instrument, is identical to the aforementioned predetermined geographical reference frame [g] except that its abscissa axis is oriented along the "free azimuth" heading of said at least one inertial measuring instrument.

[0136] The definition of this platform reference frame [p] allows us to overcome the consequence of the convergence of the meridians, so that the angular reference does not depend on the longitude, its deviation from geographic North (function of the variations in longitude) is known as the wander angle (wa).

[0137] Figure 4 also shows the aircraft's reference frame 66, also referred to hereafter as reference frame B, having an origin corresponding to that of the reference frame of said at least one inertial measuring instrument, an abscissa axis in the aircraft's lubricated plane, following the roll rotation axis 68, oriented forward, an ordinate axis in the aircraft's lubricated plane, following the pitch rotation axis 70, oriented towards the right wing and an altitude axis so that the aircraft's reference frame 66 is orthonormal right-handed (i.e. downwards if the aircraft's attitude is zero).

[0138] Such a reference point 66 of the aircraft is used by said inertial measurement unit to provide the angles of:

[0139] - Roll (cp) around the roll axis 68 corresponding to the aircraft's longitudinal axis, positive if banked to the right, - Pitch (0) around the pitch axis 70, notably defined with respect to the horizontal plane and positive when nose-up, and

[0140] - yaw i around the yaw axis 72, notably defined in relation to the Earth's geographic north, and positive towards the East.

[0141] defining the aircraft's attitude using the bank angle 1 said radar defined in relation to the aircraft's faith plane and in particular positive if the radar points downwards.

[0142] For the post-integrations to be effective, the projection step 36 in Figure 2 allows all the elementary radar measurements acquired during the period 8t to be brought back krelating to the same point of the scene targeted by the radar (M, determined by jo(0(n)[ r ( n) ], a( / )(n)[ r ( n) ] and optionally ^(e)(n)[ r ( n) ] (which, in case of unavailability, can be replaced by the distance measurement Æ( n )[r(n)] on the ground and the pitch angle) in the same frame of reference, this frame, called the unified frame of reference, taking into account the rotational and translational movements of the aircraft during the period 8t k , and being updated according to the last elementary radar measurement acquired during period 8t k , at time t k

[0143]

[0144] More precisely, such harmonization, by said projection 36, is achieved by reducing the elementary radar measurements of the instants denoted (n) (i.e. with an index n as illustrated previously on the scale 52, and expressed in the radar frame [r(n)] associated with the instants denoted (n) (i.e. with an index n, such that t n_first ,..., t n ,..., t n_last), in the platform frame associated with the times denoted (n) (i.e., index ri) denoted [p(n)], then in the platform frame [p( )] associated with the times denoted ( ) (i.e., index k of the timescale 56 represented in relation to Figure 3), considering that the change of frame from [p(n)] to [p( )] is simply a translation determined by integrating the inertial velocity measurements in the frame [p] (or, alternatively, velocity measurements obtained via radar or GNSS). Indeed, the impact of the Earth's curvature on the orientation of the vertical axis is less than 3 / 1000° (at 120 m / s over 2” (i.e., two seconds)).

[0145] Note that, since elementary radar measurements are a function of quantified values ​​in terms of distances and circulars, the allocation of these measurements to the appropriate distance (i) and circular (j) "boxes" (in the unified frame) is a function of, or weighted according to, the proximity of these measurements to the adjacent "boxes" of the unified frame.

[0146] The Cartesian coordinates of a point M in the scene targeted by the radar, expressed in [r(n)], are as follows according to equation (1):

[0147]

[0148] with here O r the origin of the radar marker [r].

[0149] For projection 36, reducing the elementary radar measurements to the times denoted (n), several coordinate transformation matrices are used, namely:

[0150] - a first change-of-reference matrix, hereafter called P2B, allowing the transition from the platform reference frame (namely the reference frame of said at least one inertial measuring instrument) to the aircraft reference frame 66 illustrated by figure 4, at time (n):

[0151]

[0152] with, for the sake of simplified notation: ip = ip(n), 9 = 9(n), <p = <p(n), les mesures inertielles obtenues précédemment via ladite étape de synchronisation 34, et ce en considérant un ordre des rotations dans ce changement de repère P2B, avec cap (ip), puis tangage (0), puis roulis (<p);

[0153] - a second change of reference matrix, hereafter called B2P, allowing conversely to go from the reference frame 66 of the aircraft illustrated by figure 4 to the platform reference frame (namely the reference frame of said at least one inertial measuring instrument):

[0154] B2P(n) = [P2B(n)] -1 = [P2B(n)] r, (i.e. the transpose matrix of P2B), the change-of-reference matrices (P2B and B2P) being considered available at times (n) and (k).

[0155] - a third reference frame transformation matrix, hereafter called R2B, allowing conversion from the radar reference frame to the aircraft reference frame 66 illustrated in Figure 4. This matrix, constant per installation, is defined as P2B, but with 9 = y (the radar tilt as indicated previously), and ip = <p = 0;

[0156] - a fourth coordinate system transformation matrix, hereafter called B2R, allowing conversely to go from the 66 coordinate system of the aircraft illustrated by figure 4 radar coordinate system to the B2R = [R2B] coordinate system -1 = [R2B] T , (i.e., the transpose matrix of R2B).

[0157] The use, during projection 36, within the frame change matrices, of the inertial measurement respectively associated with each of the said elementary radar measurements makes it advantageous to take into account the rotational and translational movements of the aircraft during the post-integration acquisition period. Using the aforementioned frame changes, the transition through the frames [p(n)] then [p(c)] is expressed in the following form:

[0158]

[0159] assuming that the orientation of [p] is constant for 8t k and that the radar installation in the aircraft is fixed

[0160]

[0161] alternative title, by differentiation of the positional quaternion:

[0162]

[0163] considering P(n)2P(c) = I (with I the identity matrix), the orientation of [p] being constant for 8t k ).

[0164] In light of the above, it is then possible to express (i.e., reduce) all the elementary radar measurements of the period 8t k in [r(k)] associated with the last elementary radar measurement (of the period 8t k ) at time t k -

[0165]

[0166] and this by using for each of the elementary radar measurements of index n (i.e., each elementary radar measurement acquired between times t n_first and t n_last illustrated by figure 3).

[0167] Starting from the Cartesian coordinates of the elementary measurement points (M) of the radar in [r(n)], as expressed previously by means of equation (1), it is indeed possible to determine the elementary measurement points M of the radar in [r(k)].

[0168]

[0169] starting from the expression of M in

[0170]

[0171] by reverting the expression for the Cartesian coordinates of M applied in the frame [r(k)] and considering that ||o r ( / c)M(i,j)(n || = p(i)(n)[p( k)] = ρ(i)(n) [r(k)] , (i.e., for any frame of reference ([r] or [p], in (k)), the distance of the measurement brought back to O r ( / c) is retained).

[0172] The equation on the line zM(i,j,e)(n) [r(k)] allows us to deduce the elevation

[0173]

[0174] , and one or the other of the other two lines {x M or y M ) then allows us to determine the circular «O')(«)[r(fc)]- The set of elementary radar measurements acquired during the time period 8t k , according to the radar grid squares i, j and e, are then all brought back and available in the same radar frame [r(k)] and therefore available from O r ( / c), which allows the "speed migrations" to be erased, whether in translation or rotation.

[0175] As an optional addition, as previously indicated, said projection includes at least one weighting determined according to the predetermined quantification of the measurement axes of said at least one radar. More specifically, this optional step 44 aims to allocate elementary radar measurements according to the quantification of the radar measurement axes.

[0176] To do this, the adjacent "boxes", denoted by the mathematical symbol "'", of each elementary radar measurement are first determined by simple comparisons with the quantized values ​​corresponding to the distance (i) and circular (j) "boxes" of the radar, namely p(i') and p(i r + 1) the values ​​corresponding to the distance boxes surrounding (i)( ) [r ( fe) ], and «( / ') and a(j r + 1) the values ​​corresponding to the circular boxes enclosing a( / )( ) [r ( fe) ], with p(i' + 1) > p i') and a(f + 1) > «( / ').

[0177] According to this optional addition, the distribution of each elementary measure is then weighted according to its proximity to adjacent boxes:

[0178]

[0179] «( / ')(«) [r(fc)] - «( / ") n J «( / ' + 1) - «( / ')(«) [r(fc)] Pond - — — — - 7777 —f ' Pond ^fji — - — — — - 7777 - +D a(j + 1) - «( / ') a(j + 1) - «( / ')

[0180] The table of elementary radar measurements reduced to [r(k)], a function of i, j, of each instant (n) composing the period 8t k is initialized to 0, then augmented in this way for each elementary measure (i, j) taken in [r(n)]:

[0181]

[0182] Note that as an alternative to weighting, elementary radar measurements are directly allocated to the nearest box (i', j'), after operating, or not, a formation of narrower beams in terms of circulars.

[0183] Note that the radar's range and circularity definition domain can be extended to take into account the aircraft's flight envelope during the St time period k , or even during the entire approach phase, particularly via sliding windows as described below.

[0184] The domains are typically extended by approximately the maximum translational or rotational speed multiplied by the time period St k (in terms of i' and j'). Note also that the expression of the elementary radar measurements of the period 8t k in [r ( / c)] preserves the phase of the 0(j,j)(n), therefore the Doppler estimation (based on this information) is not disturbed.

[0185] As an alternative, to the unified reference frame corresponding to the radar reference frame [r ( / c)] at the time of the last elementary radar measurement acquired over said predetermined acquisition period, the unified reference frame according to the present invention is also suitable for corresponding to the local geographical reference frame [r ( / c)] at the time of the last elementary radar measurement acquired over said predetermined acquisition period.

[0186] As an optional addition, a recalibration step of said unified reference frame is implemented according to the present invention by tracking at least one point of interest during said predetermined acquisition period and the preceding predetermined acquisition period.

[0187] More specifically, the localization of points of interest (POIs), particularly reflectors within a landing scene, or points of interest associated with the straight lines delimiting a runway, is generally implemented through specific processing techniques to improve their localization and the estimation of their amplitude. These specific techniques are generally performed at the level of the radar's elementary measurements, therefore in their original coordinate system [r(n)], and the POIs are identified by selecting measurements exhibiting a contrast (signal-to-noise ratio) greater than a threshold relative to ground clutter (equated with surrounding noise), or relative to thermal noise.

[0188] Ground clutter (from the English word "clutter," similar to surrounding noise) can be identified by analyzing (e.g., classifying by levels) the histogram of amplitudes (from radar grid measurements), these amplitudes being normalized to compensate for the effects of distance and physical antenna lobes.

[0189] By applying an amplitude deviation algorithm (i.e., quotient of the difference over the sum of the amplitudes of the boxes framing a point of interest POI), such as described for example in patent EP0017532 B1, in distances, circles and elevations of the points of interest, their location and amplitude in reception are refined.

[0190] Note that, in the case of circulars, it is preferentially classic to apply a phase deviation algorithm (antenna division and measurement of the phase deviation from the points of interest, such as described for example in the article entitled "Divide measurement estimator for single-pulse radar" by R. Fortet, Signal Processing, volume 4 of 1987.

[0191] Subsequently, we consider, (A(poi)(n), p(poi)(n), a(poi)(n), p(poi')(n)') the amplitudes and coordinates of a point of interest POI (poi) identified in the elementary radar measurements (n) (of index n (i.e., each elementary radar measurement acquired between times t n first et t n last illustrated by figure 3)), and therefore in the frame [r(n)].

[0192] For amplitude correction, these coordinates remain expressed in their original frame [r(n)], the amplitude of the points of interest POI can be classically corrected according to the antenna diagrams (in azimuths and elevations), by determining, in [r(n)], the distance of each point of interest POI from the center of the nearest angular cell, (i.e. the one presenting the highest level of contrast), by determining the attenuation by the antenna lobe as a function of this distance (e.g. amplitude reduced to a value of 0.8 instead of 1 at the center of the cell), and by dividing the measured amplitude by this attenuation.

[0193] On the other hand, according to the present invention, for tracking and re-regulation, these coordinates are then brought back into the unified (i.e. harmonized) frame [r( / c)], or alternatively [g ( / c)] as indicated previously.

[0194] The recalibration by tracking points of interest (POI) according to the present invention advantageously makes it possible to overcome inertial drifts.

[0195] Indeed, by definition, inertial aids (i.e. inertial measurement instruments) ensure good short-term stability (corresponding to the time horizon of post-integrations), but, without hybridization, they classically drift over time.

[0196] The tracking of points of interest (POI) proposed according to the present invention then makes it possible to recalibrate the unified frame [r ( / c)] (used according to the present invention as a reference for post-integrations) according to the observations of the POIs (and therefore not only by the accumulation of inertial variations).

[0197] Tracking consists of searching, in the following cycle, for each point of interest (POI) identified in the elementary radar measurements of a measurement cycle associated with the instants (n) composing the period 8t k(i.e., with an index n, such that t n_first ,..., t n ,..., t n_last ), and this within the perimeter (i.e., the sum) of the uncertainties of the inertial and radar measurements, in terms of distances, circular paths, and elevations. The appropriate reference frame for this search is preferably [r ( / c)] (or alternatively [g ( / c)] as previously indicated).

[0198] The uncertainties of radar and inertial measurements are a function of their standard deviations. The multiplicative factor applied to the standard deviations is determined according to the objectives (i.e., the probabilities) of detection and false alarms.

[0199] Points of interest (POIs) found throughout (or during a predetermined percentage) of the post-integration period 8t k are called "tracked POIs".

[0200] The coordinates of the same tracked point of interest (POI) (relative to each set of elementary measurements associated with the times denoted (n) (i.e., with an index n, such that

[0201]

[0202] last) are brought back into [r( / c)], then averaged to establish the best estimate of the coordinates of this point of interest POI tracked in [r( / c)].

[0203] For example, for this averaging, the barycenter weighted by the inverse of the standard deviations is preferentially used.

[0204] Similarly, the amplitude of such a tracked point of interest (POI), relative to the location period 8t k , for example, is estimated by averaging or using the center of gravity of the elementary amplitudes

[0205]

[0206] When a point of interest (POI) is tracked, its position expressed in [r(k)] is brought back to the next cycle in [r(k + 1)] to serve as a reference for its search in the elementary measurements of the period δt k+1 (in place of the first detection of this point of interest (POI)).

[0207] It should be noted that in the case of human piloting, aircraft localization is relative to the display of standardized patterns, such as approach and runway edge lights, which have standardized patterns and spacing. This recalibration improves the accuracy, jitter, and contrast of this display. Indeed, taking these geometric properties into account allows for refining the inertial recalibration by searching for these patterns within all points of interest (POIs), with uncertainties that depend on those of the measurements, and then by seeking the highest possible probability of the pattern's location.

[0208] In the case of an autopilot, the initial location is to be established by means of one of the techniques described in the patent applications in the name of the Applicant, FR 3 103 178, FR 3 152 600 or FR 3 152598, using the harmonized elementary radar measurements according to the present invention (i.e. projected in the same unified frame) and / or the points of interest (POI) located as previously described.

[0209] More specifically, using harmonized elementary radar measurements and points of interest (POIs) located as previously described, the coordinates of tracked POIs in [r(k + 1)] are predicted from their coordinates in [r ( / c)] and according to inertial measurements, and then compared to the coordinates from the radar measurements (same tracked POIs and same reference frame [r(k + 1)]).

[0210] The centroid of the discrepancies thus identified by this comparison (i.e., across all tracked points of interest (POIs), weighted according to the inverse of the standard deviations of each tracked POI) reflects the inertial drift and allows [r(k + 1)] to be recalibrated accordingly. In other words, the radar grid [r(k + 1)] is recalibrated in distances, azimuths, and elevations according to the centroids of these discrepancies.

[0211] Differences in distances advantageously identify accelerometric drifts, and differences in azimuth and elevation identify gyroscopic drifts.

[0212] Preferably, especially for determining the standard deviations of the location of tracked points of interest (POIs), the registration is performed using a weighted least squares (WLS) function or by filtering (filtering of a type belonging to the group including at least:

[0213] - a Kalman filter;

[0214] - Bayesian filtering;

[0215] - particle filtration,

[0216] and this by considering that [r( / c)] is initialized according to the grid of the first radar measurements, and that the covariance matrix is ​​initialized according to the resolutions of the Radar grid.

[0217] Note that the estimation of the standard deviation of radar measurements, necessary for the POI tracking functions and inertial state recalibration, is classically calculated using the Woodward formula (circular) and Bernoulli formula (distances):

[0218]

[0219] with ResD the distance resolution, θ 3dB the circular aperture of the global antenna, SNR, the signal-to-noise ratio of the measurement, p the number of integrated elementary measurements, k the slope of the deviation function.

[0220] Note

[0221]

[0222] As an optional supplement, the calculation of residual harmonization errors is specific to relying on the technique described in the patent application in the name of the Applicant FR 3 091 777.

[0223] A person skilled in the art will understand that the invention is not limited to the embodiments described, nor to the particular examples of the description, the embodiments and variants mentioned above being capable of being combined with each other to generate new embodiments of the invention.

[0224] The present invention thus makes it possible, by relying on inertial information, to reduce or even eliminate the degradation of the measurements of a radar, in particular pulsed, or even continuous wave, due to the movements of the aircraft to be located and / or detected, and this at a lower cost since the order of magnitude of the performance of the required inertial measuring instrument is limited to gyroscopic drifts of the order of 50 to 150° / h (i.e. an allocation of one tenth of the performance in the case of an integration over 1” (i.e. one second), TTA at 2” (i.e. two seconds)).

[0225] Thanks to this improved accuracy, it is then possible, within the framework of an aeronautical application, to more easily achieve the performance of angular measurements throughout the entire flight domain of an aircraft, with regard to the normative aeronautical requirements for approach and landing assistance.

[0226] Harmonizing radar measurements with inertial measurements allows for improved efficiency of the localization solution with post-integration (i.e., inconsistent integration), particularly when used in aeronautics to descend below 200 feet above ground level and land.

[0227] Furthermore, the point of interest (POI) tracking recalibration according to the present invention advantageously compensates for inertial drifts.

Claims

DEMANDS 1. A method (28) for three-dimensional localization of an aircraft, said aircraft being equipped with at least one radar and at least one inertial measuring instrument, said method comprising at least the following steps: - obtaining (30), via said at least one radar, elementary radar measurements acquired at measurement times distinct from a predetermined acquisition period, each measurement being acquired in a predetermined radar reference frame associated with said radar and associated with said time of its acquisition; - obtaining (32), via said at least one inertial measuring instrument, inertial measurements acquired over said predetermined acquisition period; - synchronization (34) of said elementary radar measurements and of said inertial measurements by estimation of the inertial measurements corresponding to the distinct measurement instants representative of each of said elementary radar measurements; - projection (36) of each of said elementary radar measurements in the same unified frame associated with the same predetermined instant of said predetermined acquisition period, using, for each of said elementary radar measurements, the estimated inertial measurement associated with the same acquisition instant as said elementary radar measurement considered; - three-dimensional localization (38) of said aircraft by inconsistent integration of said projection.

2. Method (28) according to claim 1, wherein said estimation (40) comprises a polynomial interpolation taking into account, for an elementary radar measurement acquired at an acquisition time, the first inertial measurement acquired after said acquisition time and a predetermined number of previous inertial measurements.

3. Method (28) according to claim 1 or 2, wherein said predetermined instant of said predetermined acquisition period is the acquisition instant of the last elementary radar measurement acquired on said predetermined acquisition period.

4. A method according to any one of the preceding claims, wherein said projection comprises at least one weighting determined according to the predetermined quantification of the measurement axes of said at least one radar.

5. A method according to any one of the preceding claims, comprising a re-registering step of said unified reference frame by tracking at least one point of interest during said predetermined acquisition period and the preceding predetermined acquisition period.

6. A method according to claim 5, wherein said registration is performed using a weighted least squares function or by filtering of a type belonging to the group comprising at least: - a Kalman filter; - Bayesian filtering; - particle filtration.

7. Computer program comprising software instructions which, when implemented by a programmable electronic device, implement a method for three-dimensional localization of an aircraft according to claims 1 to 6.

8. Three-dimensional aircraft localization device, said aircraft being equipped with at least one radar and at least one inertial measuring instrument, characterized in that said device comprises: - a first acquisition module (12) configured to obtain, via said at least one radar, elementary radar measurements acquired at measurement times distinct from a predetermined acquisition period, each measurement being acquired in a predetermined radar reference frame associated with said radar and associated with said instant of its acquisition; - a second acquisition module (14) configured to obtain, via said at least one inertial measuring instrument, inertial measurements acquired over said predetermined acquisition period; - a synchronization module (16) configured to synchronize said elementary radar measurements and said inertial measurements by estimating the inertial measurements corresponding to the distinct measurement instants representative of each of said elementary radar measurements; - a projection module (18) configured to project each of said elementary radar measurements into the same unified frame associated with the same predetermined instant of said predetermined acquisition period, using, for each of said elementary radar measurements, the estimated inertial measurement associated with the same acquisition instant as said elementary radar measurement considered; - a three-dimensional localization module (20) configured to localize said aircraft by inconsistent integration of the projection provided by the projection module.

9. Aircraft landing aid system comprising at least one radar and one electronic three-dimensional positioning device according to claim 8, said system being configured to implement said at least one radar and said device using a sliding implementation window.

10. Aircraft comprising at least one inertial measuring instrument, and a landing aid system according to claim 9.

Images