Method for determining vehicle navigation data by correlating sonar or radar data and hybridizing it with an inertial measurement unit, and a system capable of implementing such a method

Abstract

Method for determining the navigation data of a vehicle by correlation of sonar or radar data and hybridization with inertial unit, and system capable of implementing such a method. The present invention relates to a method for determining the navigation data of a vehicle having an inertial unit (6), an antenna (7) and a processing module associated with the antenna (7).The method comprises: a time alignment step (22), in which the time offset (TO) of the inertial navigation data (D0_INS) relative to the antenna navigation data (D1_SAS) is corrected; a calculation step (24) of hybrid antenna navigation data (D2_SAS) from the antenna (7) and the aligned inertial navigation data (D2_INS); and a bias correction step (26), in which an error in the aligned inertial navigation data (D2_INS) inherent to the inertial navigation system, called bias, is corrected based on the hybrid antenna navigation data (D2_SAS). Figure 6 for the abstract.

Description

Title of the invention: Method for determining the navigation data of a vehicle by correlating sonar or radar data and hybridizing it with an inertial measurement unit, and a system capable of implementing such a method

[0001] The present invention relates to a method for determining the navigation data of a vehicle, of the type having an inertial navigation system providing navigation data and sonar or radar equipment providing navigation data. It also relates to a system capable of implementing such a method and a method for imaging underwater targets comprising such a method.

[0002] The invention is primarily developed in the underwater domain, with a sonar sensor as an application case, but the problems of an airborne or space-based platform equipped with radar are exactly the same, with identical equations. In fact, the main algorithms are the same in both communities.

[0003] In the fields of underwater prospecting, bathymetry, and mine warfare, sonar can be used to obtain an image of a target encountered by a vehicle operating beneath the sea surface during its navigation. In a more advanced mode, to construct a three-dimensional image, the sonar acquires several views of the target from different angles, which are then superimposed to reconstruct the target. Another advanced mode consists of reprojecting the sonar images onto a geographic coordinate system and summing, coherently or incoherently, the received signals pixel by pixel. Currently, this registration is generally performed by image processing operations carried out after the fact: on the one hand, this does not necessarily extract the full possible accuracy from the recorded data; on the other hand, the registered navigation data is not available in real time within the vehicle.

[0004] By knowing the position of the vehicle at the times corresponding to the different shots, the formation of the 3D image can be automated. Devices are known that allow the navigation trajectory, i.e. the position of the vehicle at any time, to be estimated: inertial navigation systems and sonars can be cited in particular.

[0005] Several so-called hybridization processes are also known in which the inertial navigation data are corrected using data from the sonar. A well-known first process is called hybrid P2C2 (hybridping-to-ping cross-correlation). This algorithm can be seen as an improvement on the P2C2 (ping-to-ping cross-correlation) algorithm using sonar data alone.

[0006] The "Displacement of Phase Center Algorithm" (DPCA) method described in the seminal article by RW Sheriff, "Synthetic aperture beamforming with automatic phase compensation for high frequency sonars", Proceedings of the 1992 Symposium on Autonomous Underwater Vehicle Technology, IEEE, pp. 236-245, 1992, is a special case of P2C2 that is well known and used in practice.

[0007] The hybrid P2C2 is described in document WO99 / 18452 and in D. Billon and F. Fohanno, "Two improved ping-to-ping cross-correlation methods for synthetic aperture sonar: theory and sea results", OCEANS '02 MTS / IEEE, Biloxi, MI, USA, 2002, pp. 2284-2293 vol.4.

[0008] One difficulty among others in the application of hybrid P2C2 algorithms arises from the fact that sonar data and inertial measurement unit data are not always perfectly synchronized in time.

[0009] In general, P2C2-type algorithms are insufficient to obtain a sufficiently accurate measurement of the relative antenna displacement. Various hybridization techniques are known to those skilled in the art to improve the accuracy of this measurement. Reference may be made to the article by RE Hansen, TO Sæbp, K Gade, and S Chapman, "Signal Processing for AUV Based Interferometric Synthetic Aperture Sonar," in Proceedings from Oceans 2003 MTS / IEEE, San Diego, CA, USA, September 2003, for an overview of possible strategies. Two categories of hybridization are distinguished: tight hybridization and loose hybridization.

[0010] The first category, called tight hybridization, is a unique optimal filter (of the Kalman, extended Kalman, particle filtering, or other equivalent methods) whose state vector is the error in the vehicle's position and attitude, as well as for the various parameters involved in the measurement system (bias, misalignments, etc.). This filter takes as input the inertial increments (velocities and rotations) measured by an inertial measurement unit (IMU), pressure sensor measurements for pitch, but also (conventionally) velocity measurements taken by a Doppler log, and finally the data estimated by P2C2 (the specific variant of P2C2 is not relevant here). This filter uses the equations for the mechanization of an IMU and estimates, as accurately as possible, the position and attitude error, as well as all the errors of the overall measurement system, at each instant.

[0011] The second category, called loose hybridization, is a two-stage filter. A first stage, typically implemented in an inertial navigation system with hybridization capabilities such as the Exail Phins system, is, as in the case of tight hybridization, an optimal filter using all the information available to the system; except that, unlike tight hybridization, the measurements from the P2C2 are not used. This first filter provides the user, at the output of the control unit, with an estimate of the vehicle's position, attitude, and velocity (translational and rotational), as well as an estimate of the variance-covariance matrices for these parameters. Internal errors (biases, misalignments, etc.) are not provided to the user. A second optimal filter, downstream of the first, takes into account these position and attitude estimates and their time derivatives, as well as the data from the P2C2, in order to arrive at a hybrid result.

[0012] The tight hybridization approach is more precise than loose hybridization but requires access to the raw inertial increments, which is not always possible for various technical and regulatory reasons. The second method is somewhat cruder but simpler to implement.

[0013] The aim of the invention is therefore to propose a method for determining the navigation data of a vehicle which allows for increased navigation accuracy, and consequently for improved quality of automatic reconstruction of 3D images of targets, through a particular loose hybridization strategy, all without involving a more efficient inertial navigation system whose cost and size would not be compatible with the applications targeted.

[0014] To this end, the invention relates to a method for determining the navigation data of a vehicle equipped with an inertial measurement unit (IMU) providing IMU navigation data, an antenna providing raw antenna data, and a processing module associated with the antenna and configured to determine antenna navigation data from the raw antenna data. The method comprises correcting the IMU navigation data from the antenna navigation data. The method includes: • a time alignment step, during which the time offset of the inertial navigation data relative to the antenna navigation data is corrected, providing time-aligned inertial navigation data; • a step of calculating hybrid antenna navigation data, from the antenna and the calibrated inertial navigation data; and • a bias correction step, during which an error in the recalibrated inertial navigation data inherent to the inertial navigation system, called bias, is corrected based on the hybrid antenna navigation data, providing recalibrated and debiased inertial navigation data, forming the vehicle's navigation data

[0015] Thanks to the invention, the time delay that may exist between the data from the two measuring devices, namely the inertial measurement unit and the sonar or radar antenna, The inertial measurement error, which originates from the hardware and can vary over time, is corrected. This temporal realignment allows for a reliable correction of the cumulative positioning bias inherent in the inertial navigation system, using measurements from the hybridized antenna itself. Following these corrections, navigation accuracy is improved, notably enabling greater precision in the automatic reconstruction of 3D target images.

[0016] According to other advantageous aspects of the invention, the method for determining the navigation data of a vehicle comprises one or more of the following features, taken individually or in all technically possible combinations:

[0017] - the process is such that: • the corrected bias of the recalibrated inertial navigation data is a velocity bias; • Recalibrated and debiased inertial navigation data includes recalibrated and debiased inertial velocity data; and • the process includes a trajectory reconstruction step during which a vehicle trajectory is calculated by time integration of the recalibrated and unbiased inertial measurement unit velocity data, providing the vehicle trajectory, forming the vehicle navigation data;

[0018] - the antenna is a synthetic antenna sonar antenna and the processing module determines antenna navigation data using an antenna displacement estimation algorithm by correlation;

[0019] - the process is such that: • The recalibrated inertial navigation data includes recalibrated attitude data; and • Hybrid antenna navigation data is obtained by a hybrid antenna displacement estimation algorithm involving recalibrated attitude data;

[0020] - the time alignment step includes: • an expression of inertial navigation data in an antenna navigation data space, providing modified inertial navigation data; • an estimate of the time lag by correlation between the modified inertial navigation data and the antenna navigation data; and • a subtraction of the time offset in the inertial navigation data, providing the recalibrated inertial navigation data;

[0021] - the time alignment step includes filtering the navigation data from inertial navigation system, modified inertial navigation data and / or antenna navigation data;

[0022] - the bias removal step includes: • an expression of hybrid antenna navigation data in the inertial navigation data space, providing modified hybrid antenna navigation data; • an estimation of the bias based on modified hybrid antenna navigation data and recalibrated inertial navigation data; and • a removal of bias in the recalibrated inertial navigation data, providing recalibrated and unbiased inertial navigation data;

[0023] - the bias removal step includes filtering the navigation data hybrid antenna or modified hybrid antenna navigation data;

[0024] - the vehicle is an underwater vehicle, and the antenna is a sonar antenna.

[0025] The invention also relates to a system for determining the navigation data of a vehicle, comprising a means for receiving navigation data from an inertial measurement unit and a processing module associated with an antenna and capable of implementing a method for determining the navigation data of a vehicle as defined above.

[0026] It also relates to a method of three-dimensional underwater target imaging using a multi-view sonar comprising the implementation of a method according to the above and a method of registering the views from the navigation data of the underwater vehicle from the first method.

[0027] 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:

[0028] [Fig. 1] [Fig. 1] is a perspective view of a vehicle comprising a system capable of implementing the method according to the invention to determine its navigation data,

[0029] [Fig.2] [Fig.2] is a diagram of an antenna belonging to the vehicle of [Fig.1] to two different moments, introducing motion parameters;

[0030] [Fig.3] [Fig.3] is a diagram of the antenna of [Fig.2] at two different times, introducing the concept of a phase center;

[0031] [Fig.4] [Fig.4] is a diagram representing the antenna of [Fig.3] and a target;

[0032] [Fig.5] [Fig.5] is a diagram representing two sets of phase centers of the antenna of [Fig.3];

[0033] [Fig.6] [Fig.6] is a flowchart representing the different steps of the process according to the invention.

[0034] Consider a vehicle 1. The vehicle 1 is advantageously a submersible vehicle 1, as shown in [Fig. 1]. In an alternative not shown, the vehicle 1 is an airborne or space-borne platform.

[0035] The underwater vehicle 1 shown in [Fig.1], when in navigation, is partially or totally submerged and navigates at a distance from a seabed 3.

[0036] The underwater vehicle 1 is used in the context of underwater prospecting, wreck searches, or mine warfare, which encompasses any operation related to underwater mines. In particular, the underwater vehicle 1 is capable of detecting a target 5 encountered during its navigation, notably using target imaging techniques.

[0037] The underwater vehicle 1 has an inertial navigation system 6, located inside the vehicle, and at least one antenna 7 positioned outside the underwater vehicle 1. It preferably has two antennas 7, located respectively on the port and starboard sides of the underwater vehicle 1. The underwater vehicle 1 also includes a processing module 9 associated with the antenna 7 and a computer 8 to which the inertial navigation system 6 and the processing module 9 are connected for processing the received signals.

[0038] The term "inertial unit" here refers both to a standalone inertial unit and to an inertial unit already hybridized in a known way with external devices, for example with a Doppler log, a pressure sensor and a computer embedded in the inertial unit performing a process of fusion of inertial and Doppler information.

[0039] The inertial navigation system provides, in a known manner, inertial navigation system D0_INS data comprising: - The position of the underwater vehicle 1 in latitude, longitude and immersion; - The speed of the underwater vehicle 1 in the same frame of reference as its position; and - The roll heading angles 0 and pitch angles 0, shown in [Fig.1], forming the attitudes of the underwater vehicle 1.

[0040] Inherently in the operation of an inertial navigation system, the navigation data of the D0_INS inertial navigation system suffer from a cumulative error over time, referred to as bias in the rest of the description.

[0041] In the example in [Fig. 1], the antennas 7 are sonar antennas, preferably synthetic sonar antennas. In an alternative not shown, the antennas 7 are, for example, radar antennas.

[0042] The antennas 7 extend along the length of the underwater vehicle 1, substantially parallel to a main axis of the underwater vehicle 1 corresponding to a straight-line advance of the underwater vehicle 1. Each antenna 7 consists of acoustic transmitters and receivers, equally distributed along an antenna axis generally collinear with the main axis of advance of the underwater vehicle 1.

[0043] As is known per se, the transmission and reception of acoustic signals provides raw antenna data D0_SAS. Each antenna 7 is associated with a processing module 9, configured to determine antenna navigation data D1_SAS from the raw antenna data D0_SAS. In particular, as explained later in the description, the raw antenna data D0_SAS is used by the processing module 9 to estimate the displacement of the antenna 7, and therefore of the underwater vehicle 1, as well as the bearing Gt corresponding to the angle between the principal axis X and an axis between the underwater vehicle 1 and an external reference point P. Alternatively, the raw antenna data D0_SAS is received and processed by the computer 8.

[0044] Advantageously, the antenna 7 has more than one bathymetric function, allowing a bathymetric capacity to be determined using, for example, an interferometric antenna.

[0045] Advantageously, the antenna 7 has more than one so-called multi-aspect capability, allowing several images to be acquired along offset viewing axes.

[0046] Antenna 7 is also shown at different times in Figures 2 to 5, introducing different parameters useful for the description below.

[0047] The method for determining navigation data of the underwater vehicle 1, implemented by the computer 8 under the control of one or more suitable computer programs, uses a hybrid P2C2 algorithm. This algorithm can be seen as an improvement on a P2C2 algorithm using antenna data alone, which should be briefly described before describing the hybrid P2C2.

[0048] The following description is given with reference to Figure 3. The sonar antenna 7 consists of a transmitter whose position at time t = nTR of the emission of the nth pulse is denoted Xn (with TR the pulse repetition period), and G receivers typically aligned and equally distributed at a position (1 < QT < G)- We consider a target 5 located at a range rnp = Eq + pAr of Xn, with q a certain initial, constant and known range on the order of a few meters and Ar a A certain distance step is very small compared to the maximum sonar range cT (Af is on the order of a few centimeters and cT^ / 2 on the order of a few hundred meters). During the round-trip flight time of the sonar pulse to the target 5, the antenna 7 moves, undergoing a continuous translational and rotational motion, so that the receivers of antenna 7 are no longer in the same position (1 < ÇT < G), than at the date of transmission, but at a position R& 1 < g < Q (which depends on the distance rap + r^p transmitter-target- 9th receiver and the parameters of the movement of antenna 7: the further away the target 5 is, the longer the time of flight c[rap + Fnp) is 'onS ct Plus 'antcnnc 7 has time to move). Implicitly the signal arrives at the ^'th sensor of the receiving antenna 7 at a date t = nTR + c(r0+p&r + rgp) = tnp In the equations that follow, the index pair (n, p) implicitly corresponds to this date tnp-

[0049] We denote C^p a phase center of the 9th transmitter-receiver pair (Xa, Rnp)' This phase center corresponds to a hypothetical position of antenna 7 if it had not moved between transmission and reception. It can be shown that this is the midpoint of the segment. Implicitly, C^p therefore depends on tnp, and thus on the distance to the target 5 and the movement of vehicle 1. However, to a first approximation, we can consider that, with ^np fixed, the -[Cap} are aligned on a straight line parallel to the axis of antenna 7 at the average signal reception time, that is, with the points

[0050] The image plane is now defined as follows, with reference to Figure 4. First, the antenna 7, at its position and attitude at time tnp, must be equipped with an antenna frame ( ) such that r ? is along the longitudinal axis [Onp, Pnp, Çnp) tnp of antenna 7, and in a main pointing direction essentially orthogonal to a surface of antenna 7. We denote by pa the unit direction vector going from the origin On of antenna 7 to the target 5 at tnp. We then denote by VnP the unit vector equal to the normalized version of the projection of p* onto the axis (q p-*)• The plane defined by these two vectors p^ and VnP and passing through O, is the image plane corresponding to target 5. The grazing angle Pnp of the image plane is such that ^np = ( 0 cosy siny jt (knowledge of this grazing angle requires a measurement, classically carried out by interferometry).

[0051] Considering now two consecutive sonar pulses emitted at (n - 1) Tr and nTr, we thus have two sets of phase centers, represented in Figure 5, for the same target 5 located at a distance from the emitter at the nth pulse. These two sets are denoted J and {Cnp}' with 1 — — G- The P2C2 algorithm aims to to determine, by correlating the received signals, the sub-antenna portions, that is to say / ^max\ and / ^min ç,max\ lc|s that |cs signals received at phase centers KX 1 < GX < g < GX < g} '{Cnp1 — Gnp1 — ff — Gnp* — G} correlate best, which simultaneously allows us to determine the relative position of these sub-antennas (at the phase center level), and the position of antenna 7 at the reception times. In Figure 5, the index of the sub-antenna center at the reception of pulse n-1 is denoted j, and its counterpart for pulse 11 is denoted (Ie). We have and DP 9^ ""2^...... G min । Ji +G" 2

[0052] The P2C2 algorithm provides an estimation of parameters of Ln, ^n, Tn movement maximizing the correlation of the received signals between the two sub-antennas {cX 1 s GX < g < G“S < g} ■ { Clp, 1 < ^g^ GX< G} : defined as above, and more formally as the vector orienting the bisector axis of the angle dg di dg Y located in the ^\KnAp-Kn.ip, rtnp-K-np I image plan and pointing towards the front of antenna 7, that is to say roughly in the direction of the antenna's translation: is therefore a kind of $n,p average of r * and r > ; re the vector orthogonal to je , in the image plane, and oriented towards the Ç n,p direction of transmission of antenna 7 (to starboard) is therefore a kind of average of p^p and rnp; (y—* . , a deviation of abscissa along the principal axis ?„P cos— between two sensors maximizing the cross-correlation between two pings ^np successive; Tnp — a time gap, CTnp NF corresponding to an ordinate difference along the axis between two sensors maximizing the cross-correlation between two successive pings and where T^p is an additional near-field term whose details are not useful for understanding the invention; the rotation allowing the transition from the orientation of the physical antenna 7 at the reception of ping H-1, to the orientation of the physical antenna 7 at the reception of ping n in the image plane defined as which also corresponds to the rotational movement to go from the axis of J 1, to

[0053]

[0054]

[0055]

[0056]

[0057] that of the JJs L ^n,p J • The motion parameters Ln, pn, Tn are represented in [Fig.2]. Formally, this P2C2 process can be seen as the resolution of: (L n . K) = argmax JJ T)s n ( ^-L, tt-)- L, p, T (The) with sn the signal received for pulse 11 and $ni the signal received for pulse n-1. Between the dates of reception of two pulses 11 and n + 1 relating to a target 5 located at a reference range rnp, vehicle 1 undergoes a rotation increment noted o ' — / f . -, , , \ t (components in radians) L2np- {toçpp, Wppp, COçpp) F in the antenna frame / -r—Y The rotation operator denoted H (Emp, Dp, Pap> >>np) which transforms the coordinates of an arbitrary vector x expressed in the frame into the coordinates in the frame p* is with, to the first order ~ X + QnP X x- or even matrix-based: Mppp ' H(x) = ~^p,np 1 Mpjip ■ ^2ppp

[0058] Angle P is therefore formally defined as follows: 100591

[0060] By setting y—* _ L nn) it comes with an excellent approximation:

[0061] Kp = - ^P-siny2p = fl(t^ (E2)

[0062] The angle is counted positively in the trigonometric direction (resp. anti- trigonometric) for a sight to port (resp. starboard). This equation (E2) links p" Pn to two gyrometric parameters which are and ^pji. Similarly, it can be shown that there is a link between Lnp, by nonlinear equations, theoretically linking Lnp, Tnp to the inertial increments of rotation denoted (components in radians) but also of velocity ÔVppp ÔVç ) f in the antenna frame (npf Pnp' Cup) [00631 Ciav7p <e3) [°064l =

[0065] The exact expression of (E3) and (E4) is complex but can be determined by elementary geometric considerations known to those skilled in the art. Their exact form is unnecessary for understanding the invention. These equations are assumed to be numerically invertible by various algorithms known to those skilled in the art.

[0066] Furthermore, it is clear that these equations (E2, but also E3, E4) depend on the index pair n'P, meaning that the estimation process is carried out densely over time tnp = IlTR + C ( Co + pAr + Tnp ), with & varying to cover all possible target ranges. The data arrival rate Lnp, ^np> Pnp is therefore much greater than the repetition period of the TR pulses (several tens of hertz).

[0067] In the case of a multi-aspect sonar antenna and with the notations introduced previously, the processing module 9 associated with the antenna 7 makes it possible to estimate two triplets on two image frames that are off-point with respect to the image frame main function and to generate sonar images in the associated landmarks.

[0068] The "Displacement of Phase Center Algorithm" (DPCA) method is a special case of P2C2 and proceeds by an approximation of equation (El) assuming that an estimate of Lnp is known because it is estimated elsewhere as for a log at

[0069] correlation ; that the spatio-temporal correlation is achieved by several temporal cross correlations of the different sensors composing the antenna, and that the pair ( ) is °btenu P31 linear regression. The hybrid P2C2 directly estimates 'g' from gyrometric data. The rotation increment is measured in the inertial measurement unit frame 6, then transformed into the antenna frame so that the coordinates of q* y are ( OJpni f (in radians). We then estimate directly using equation (E2). The rest of the algorithm then consists of estimating and by a simplification of (El) written:

[0070] (tc j. (r, (L n , T n )=max ff t)s n [^-ô , + Jd^dt (E5)

[0071] One difficulty among others in the application of (E2) and (E5) arises from the fact that the antenna data and the inertial measurement unit data are not always perfectly synchronized in time. Also, the estimated zT used in the P2C2 The hybrid function can be flawed because it may be time-shifted relative to the actual movement. This shift in P then also induces an additional error due to the coupling caused by (E5).

[0072] The method for determining navigation data of the underwater vehicle 1 according to the invention, which makes it possible to overcome the aforementioned limitations, is detailed below with reference to [Fig.6].

[0073] Initially, the D0_SAS antenna data is converted into D1_SAS antenna navigation data for each point 5 in the scene. This conversion 20 is advantageously performed using an antenna displacement estimation algorithm 201 between two pings based on a classic (non-hybrid) P2C2 (ping-to-ping cross correlation) inter-ping correlation described by the aforementioned equation (E1). It is advantageously complemented by an algorithm 202 providing the bathymetric capability to determine the grazing angle at each point 5 in the scene and / or the multi-aspect capability. The D1_SAS sonar navigation data for target 5 is a series that includes, for each temporal recurrence of algorithm 201: - a parameter 201 of abscissa deviation along the antenna axis between two sensors maximizing the cross-correlation between two successive pings; - a propagation delay parameter that temporally represents the displacement of a sensor along the principal axis in the rDp image plane between two successive pings; and - a p 201 parameter of rotation of the relative bearing of the target in the Pnp plane image between two successive pings, induced by the rotation of the antenna.

[0074] These parameters determine, for each iteration of algorithm 201, the estimates of the actual displacement Lnp, TnP, of the phase center in the image plane from the antenna measurements alone. This step does not involve the inertial navigation data D0_INS in order to guarantee the independence between the antenna navigation data D1_SAS and the inertial navigation data D0_INS.

[0075] In a second step, a time alignment step 22 subtracts from the inertial navigation data D0_INS a time offset DT that may exist between the inertial navigation data D1_INS and the antenna navigation data D1_SAS. This time offset DT is of hardware origin and may vary over time.

[0076] The time-recalibration step 22 advantageously includes a step 221 of expressing the inertial navigation data in an antenna navigation data space by calculating equations (E3) and (E4), a step 222 of estimating the time offset and a step 223 of subtracting the time offset.

[0077] During step 221, the inertial navigation data D0_INS (rotation increments in radians OJppp, MÇpp and velocity increments ΔV^p, ΔVp^p, ΔV^^p in meters per second) expressed in the frame / \ of antenna 7, so as to calculate, for [^np, Çnp, Pnp, ÇnpJ a particular point in the scene located at a distance Fnp = Fq + p / \F from the emitter at the time of emission of pulse n, parameters: Fnp which are estimates of Ln, Tn, from inertial measurements alone, using equations (E2, E3 and E4). We then obtain modified inertial navigation data D1_INS.

[0078] Advantageously, the D1_SAS 201', 201 and antenna navigation data At n n 201 and the modified inertial navigation data D1_INS f 221\ 221 and 221 G'" Pnp

[0079]

[0080] p 221 and 221 are then filtered by a bandpass filter with a cutoff frequency n Pn less than ° where R is the pulse repetition period. During step 222 of the time lag estimation, the time lag DT is estimated by correlation between parameters 201, p' 201) of ^np pnp D1_SAS antenna navigation data and G* parameters 221', ~ 221, ■Lnp nP rC 22 ï) modified inertial navigation data D1_INS. This Pnp The correlation step consists of finding ôp such that T 201' , T - 201, tn,p-6p 221). This term Ôp , 201 corresponds best to G221', 22^^ ' 'np-ôp np nP ?nP corresponds to the temporal gap rvr — to be estimated, which maximizes the inter-correlation between the parameters 201' 201 201) and (p 221' 221 221') Finally, step 223, which subtracts the time-shift data, acts on the inertial navigation data D0_INS by temporally shifting it by the previously estimated time-shift value DT, providing recalibrated inertial navigation data D2 INS 223' ~ 223 ~ 223 Qn * - Lnp ^np ^np Note D21_INS: Inertial measurement unit velocity data recalibrated and D22_INS recalibrated inertial navigation system attitude data included in recalibrated inertial navigation system data D2_INS.

[0081] In a third step, the antenna data D0_SAS are converted into hybrid antenna navigation data D2_SAS. This conversion 24 is performed using a so-called hybrid P2C2 antenna displacement estimation algorithm 241, similar to antenna displacement estimation algorithm 201, in which the recalibrated inertial measurement unit attitude data D22_INS and the target's crossing angle estimated by interferometry are also incorporated; this hybrid P2C2 algorithm is explained by the aforementioned equation (E5), using the term p'" for equations (E3, E4, E5). 223 The hybrid antenna navigation data D2_SAS, Pnp corresponding to the data / ^241, ^241 p 241\ s°nt more precise than the lL,a , Tn , pn I D1_SAS antenna navigation data regarding the rotation of antenna 7 between two successive pings. They are also advantageously more accurate than the direct application of equations (E2) and (E3) because the time offset DT has been corrected.

[0082] As before, the conversion 24 is advantageously complemented by an algorithm 242 providing the bathymetric and / or multi-aspect capability.

[0083] In a fourth step, a bias correction step 26 corrects the aforementioned bias present in the recalibrated inertial navigation data D2_INS, using the hybrid antenna navigation data D2_SAS and the recalibrated inertial navigation data D2_INS. The bias correction step 26 preferably consists of at least a step 261 for expressing the hybrid antenna navigation data in the inertial navigation data space and a step 262 for estimating and removing the bias.

[0084] In the following description and in a non-limiting manner, the bias is considered to be a speed bias, and the bias correction is therefore done on speed data.

[0085] During step 261 of expressing the hybrid antenna navigation data in the inertial navigation data space, hybrid antenna velocity data are calculated from the hybrid antenna navigation data D2_SAS, before being expressed as a function of the recalibrated inertial navigation data D21_INS, by a change of reference frame known to those skilled in the art. Formally, the velocity increments V2®1 are recovered from a Inversion of equations E3, E4, E5 evaluated for the D2_SAS hybrid antenna navigation data corresponding to / 241, ^241 n 24 A- The algorithm produces Um z Tn , ô modified hybrid antenna navigation data D3_SAS, here consisting of modified hybrid antenna velocity data 261 np ' which are advantageous filtered at the end of step 261.

[0086] The modified hybrid antenna navigation data D3_SAS, corresponding to A, are then used in step 262 of estimation and suppression of the bias. The velocity bias is estimated using an averaging operation between the modified D3_SAS hybrid antenna velocity data corresponding to and The inertial measurement unit velocity data recalibrated D21_INS. The bias thus estimated is then removed from recalibrated inertial navigation unit velocity data D21_INS, providing recalibrated unbiased inertial navigation unit navigation data D3_INS, here consisting of recalibrated unbiased inertial navigation unit velocity data.

[0087] Advantageously, in a final stage, a trajectory reconstruction step 28 estimates a trajectory D4_INS of the underwater vehicle 1 by time integration of the recalibrated, unbiased D3_INS inertial measurement unit velocity data.

[0088] Whether the corrected bias is a velocity bias or not, step 26 provides recalibrated, unbiased D3_INS inertial navigation data which can either form the navigation data of the underwater vehicle 1, or undergo a transformation during step 28, which is not necessarily a time integration as presented in the case of velocity bias, allowing the recovery of the desired underwater vehicle navigation data, usually the trajectory of underwater vehicle 1.

[0089] The precise determination of the navigation data of the underwater vehicle 1, in accordance with the process which has just been described, makes it possible for example to carry out a precise automatic recalibration of these different views, acquired by the antenna 7 during navigation at different points of view with respect to the target 5, in order to produce a 3D image of the target 5 by tomography.

Claims

Demands

1. A method for determining the navigation data of a vehicle (1) having an inertial measurement unit (IMU) (6) providing IMU navigation data (D0_INS), an antenna (7) providing raw antenna data (D0_SAS), and a processing module (9) associated with the antenna (7) and configured to determine antenna navigation data (D1_SAS) from the raw antenna data (D0_SAS), the method comprising a correction of the IMU navigation data (D0_INS) from the antenna navigation data (D1_SAS), characterized in that the method comprises: • a time alignment step (22), during which the time offset (TO) of the IMU navigation data (D0_INS) with respect to the antenna navigation data (D1_SAS) is corrected, providing time-aligned IMU navigation data (D2_INS);• a calculation step (24) of hybrid antenna navigation data (D2_SAS), from the antenna (7) and the recalibrated inertial navigation data (D2_INS); and • a bias correction step (26), during which an error on the recalibrated inertial navigation data (D2_INS) inherent in the inertial navigation system, called bias, is corrected on the basis of the hybrid antenna navigation data (D2_SAS), providing recalibrated and debiased inertial navigation data (D3_INS), forming the vehicle navigation data (1).;

2. A method for determining the navigation data of a vehicle (1) according to claim 1, wherein: • the corrected bias of the recalibrated inertial navigation data (D2_INS) is a velocity bias; • the recalibrated and unbiased inertial navigation data (D3_INS) include recalibrated and unbiased inertial velocity data; and • the method includes a trajectory reconstruction step (28) during which a vehicle trajectory (1) is calculated by time integration of the recalibrated and unbiased inertial velocity data (D3_INS), providing the vehicle trajectory (D4_INS), forming the vehicle navigation data (1).

3. Method of determining the navigation data of a vehicle (1) according to claim 1 or 2, wherein the antenna (7) is a synthetic antenna sonar antenna and the processing module (9) determines the antenna navigation data (D1_SAS) by means of an antenna displacement estimation algorithm by correlation (201).

4. A method for determining the navigation data of a vehicle (1) according to claim 3, wherein: • the recalibrated inertial navigation data (D2_INS) includes recalibrated attitude data (D22_INS); and • the hybrid antenna navigation data (D2_SAS) is obtained by a hybrid antenna displacement estimation algorithm (241) involving the recalibrated attitude data (D22_INS).

5. A method for determining the navigation data of a vehicle (1) according to any one of the preceding claims, wherein the time-recalibration step (22) comprises: • an expression (221) of the inertial navigation data (D0_INS) in an antenna navigation data space (D1_SAS), providing modified inertial navigation data (D1_INS); • an estimation (222) of the time offset (DT) by correlation between the modified inertial navigation data (D1_INS) and the antenna navigation data (D1_SAS); and • a subtraction of the time offset (DT) in the inertial navigation data (D0_INS), providing the recalibrated inertial navigation data (D2_INS).

6. Method for determining the navigation data of a vehicle (1) according to claim 5, wherein the time recalibration step (22) includes filtering the inertial navigation data (D0_INS), the modified inertial navigation data (D1_INS) and / or the antenna navigation data (D1_SAS).

7. A method for determining the navigation data of a vehicle (1) according to any one of the preceding claims, wherein the bias removal step (26) comprises: • an expression (261) of the hybrid antenna navigation data (D2_SAS) in the inertial navigation data space (D2_INS), providing modified hybrid antenna navigation data (D3_SAS); • an estimation of the bias from the modified hybrid antenna navigation data (D3_SAS) and the recalibrated inertial navigation data (D2_INS); and • a bias removal in the recalibrated inertial navigation data (D2_INS), providing the recalibrated and debiased inertial navigation data (D3_INS).

8. A method for determining the navigation data of a vehicle (1) according to claim 7, wherein the bias removal step (26) comprises filtering the antenna navigation data hybridized (D2_SAS) or modified hybridized antenna navigation data (D3_SAS).

9. A method according to any one of the preceding claims, wherein the vehicle (1) is an underwater vehicle, and the antenna (7) is a sonar antenna.

10. System (8) for determining the navigation data of a vehicle (1), comprising a means for receiving navigation data from an inertial measurement unit (6) and a processing module (9) associated with an antenna (7) and capable of implementing a method for determining the navigation data of a vehicle (1) according to one of the preceding claims.

11. A method for three-dimensional underwater target imaging using a multi-view sonar comprising the implementation of a method according to claim 9 and a method for registering the views from the navigation data of the underwater vehicle (1) from the first method.