Navigation method for a vehicle and associated device

Abstract

The invention relates to a navigation method for a vehicle comprising: estimation, by means of an inertial navigation unit (U_NAV) of the navigation device (DISP), of navigation data based on inertial data generated by a high-precision inertial measurement unit (IMU_HP); and correction of the navigation data by means of a fusion process, which is carried out by a tight hybridization unit (U_HYB) of the navigation device (DISP) and applied to the navigation data and positioning data of a terrestrial reference point derived from an image.

Description

Description Title of the invention: Navigation method for a vehicle and associated device Technical Field

[0001] The present invention belongs to the general field of navigation and positioning. More particularly, it relates to a navigation method for a vehicle. It also relates to a navigation device configured to implement such a method. The present invention finds a particularly advantageous, though by no means limiting, application in the implementation of navigation systems used during the approach phase preceding the landing of an aircraft. Previous technique

[0002] Inertial navigation systems (INS) are devices designed to assist in vehicle orientation. Specifically, it is common practice to use an INS as a navigation system on board a vehicle. More generally, we will use "inertial navigation" to refer to a navigation solution that utilizes data collected by an inertial measurement unit. This data corresponds to the specific force and angular velocity. However, using inertial data to implement a navigation solution requires addressing the well-known problem of drift over time in inertial navigation. Indeed, measurement errors of specific force and angular velocity are integrated over time by the inertial navigation system, leading to increasingly large errors in velocity and position.

[0003] To limit drift in inertial navigation, a state-of-the-art approach known as a "hybrid navigation solution" combines data from multiple sensors. Some solutions focus on combining inertial data with data from a Global Navigation Satellite System (GNSS). However, these solutions depend on the availability and continuity of the data provided by this satellite positioning system, which can be disrupted, for example, by jamming. Indeed, the very low power of a GNSS signal received by an aircraft makes the navigation solution vulnerable to interference, whether intentional or not. Ultimately, in the current environment, hybrid inertial-GNSS navigation solutions may no longer be reliable and accurate.

[0004] Computer vision techniques have also been considered to compensate for the drift of inertial navigation using images acquired by a vehicle. However, existing hybrid inertial-vision navigation solutions are largely dependent on the quality and availability of visual data and remain complex in terms of processing. Indeed, to correct the attitude of an aircraft estimated by inertial navigation, a fictitious point—generally a vanishing point—must be reconstructed from measurements of several reference points, such as the corners of a runway. These reference points are usually detectable without correlation only at a short distance from the runway, which is incompatible with the distance at which the approach phase preceding the aircraft's landing must be initiated.In this sense, existing hybrid inertial-vision navigation solutions do not allow for a sufficient level of integrity and availability for critical applications, such as aeronautics.

[0005] There is therefore a need for a navigation solution that improves the level of integrity and availability of existing navigation systems. Description of the invention

[0006] The present invention aims to remedy all or part of the drawbacks of the prior art, in particular those set out above, by proposing a hybrid inertial-vision navigation solution usable during the approach phase preceding the landing of an aircraft, in particular between the FAP (acronym for "Final Approach Point" and the DA (acronym for "Decision Altitude").

[0007] To this end, and according to a first aspect, the invention relates to a navigation method for a vehicle, the method being implemented by a navigation device and comprising:

[0008] - obtaining inertial data generated by a high-precision inertial measurement unit;

[0009] - an estimation, by an inertial navigation unit of the navigation device, of navigation data from the inertial data obtained;

[0010] - the reception, by a tight hybridization unit of the navigation system connected to the inertial navigation unit, of positioning data of a terrestrial reference point within an image acquired by the vehicle, said positioning data being previously calculated by an image processing unit of said vehicle; and,

[0011] - a correction of navigation data, by applying a fusion, by the tight hybridization unit, of navigation data and positioning data of the terrestrial reference point.

[0012] This corrected navigation data then allows the vehicle to be guided. This vehicle corresponds, for example, to a land vehicle (e.g., a car, a truck, a train), a boat, or an aerial vehicle (e.g., an aircraft, a helicopter, a plane, a drone).

[0013] In a particular implementation mode, only a portion of the navigation data is corrected.

[0014] As mentioned previously, the proposed solution therefore uses images of the vehicle's environment to compensate for inertial navigation drift.

[0015] "Navigation data" refers here to data relating to the position and / or movement of the vehicle, such as geographical coordinates (e.g., latitude, longitude, altitude), speed, attitude, etc. In the context of the invention, a "position" can refer to an absolute (or "global") position defined with respect to the Earth's frame of reference, or a relative (or "local") position defined with respect to a reference position.

[0016] The term "inertial measurement unit" refers to a measuring device that provides, for a plurality of measurement times, data relating to the specific force (i.e., the sum of external forces other than gravitational forces divided by the mass) and angular velocity of the vehicle. Furthermore, the term "inertial navigation unit" hereafter refers to a navigation device that integrates, over time, the specific force and angular velocity data produced by an inertial measurement unit and allows for the determination of vehicle navigation data.

[0017] In the context of the invention, this inertial measurement unit is referred to as "high-precision". In a particular embodiment, this inertial measurement unit is of "navigation" grade. A navigation-grade inertial measurement unit differs from a standard inertial measurement unit in its accuracy, stability, and reliability. Indeed, a navigation-grade inertial measurement unit typically exhibits a gyroscopic drift class of 10 -2degrees per hour, and a position error drift class of 1 nautical mile per hour. In contrast, standard inertial measurement units typically exhibit a gyroscopic drift class of a few degrees per hour, and a position error drift class of several tens of nautical miles per hour, which very quickly renders pure inertial navigation (without hybridization) unusable. As discussed in more detail below, this high-precision inertial measurement unit is specifically configured to provide angle and velocity increments as output.

[0018] In the context of the invention, thanks to the quality of the navigation data determined by the high-precision inertial unit, the positioning data from a single terrestrial reference point is sufficient to correct the navigation data. Therefore, modeling correlations between several terrestrial reference points and / or constructing a fictitious reference point—such as a vanishing point—from a plurality of terrestrial reference points is unnecessary. Furthermore, in a particular embodiment, absolute geographic positions are manipulated, and knowledge of the geometric properties of the runway, such as its width, is then unnecessary.

[0019] A "tight hybridization unit" is defined as a tightly coupled hybridization (also sometimes called a "fusion" unit). Generally, in the field of hybrid navigation, coupling techniques are defined according to several categories, including loose and tight coupling. When a navigation system is considered to have tight coupling, it refers to the fact that the input data to a global filter corresponds to raw measurements. The concept of "tight coupling" is defined in contrast to that of "loose coupling," in which a first filter (or estimator) is used to process initial input data, and the output of this first filter is combined with second data within a second integration filter. In loose coupling, the output of the first filter (and therefore the input of the second) corresponds to navigation data.

[0020] Compared to existing navigation solutions, the proposed navigation solution, by using a high-precision inertial measurement unit combined with a tight hybridization method, improves the integrity and availability of the vehicle's navigation system while accurately determining vehicle navigation data. Specifically, the proposed navigation solution offers higher levels of availability and integrity than existing hybrid inertial-vision navigation systems.

[0021] Indeed, the availability of the navigation device is no longer conditional on obtaining measurements from several visual reference points, which are generally only observable without correlation at a reduced distance.

[0022] Generally speaking, the steps of a process should not be interpreted as being linked to a notion of temporal succession.

[0023] In particular modes of implementation, the navigation method may further include one or more of the following characteristics, taken individually or in all technically possible combinations.

[0024] In particular embodiments, the vehicle further includes a satellite positioning system, and the method is implemented when said satellite positioning system is unusable.

[0025] In certain implementation modes, not all navigation data determined by the high-precision inertial measurement unit is corrected.

[0026] In particular, the navigation data is not corrected by the tight hybridization unit.

[0027] Indeed, in known navigation solutions, low-precision inertial measurement units are used, and the biases of these sensors must then be corrected using data from other sensors.

[0028] In certain implementation modes, the estimated navigation data includes a vehicle attitude that is directly used by a vehicle guidance module (CMD), without being corrected. In other words, a correction to this attitude is not determined by the tight hybridization unit.

[0029] In specific implementation modes, the process further includes the following steps, implemented by the image processing unit:

[0030] - image acquisition by the vehicle or by an image capture device on board that vehicle;

[0031] - detection of the terrestrial reference point in the image acquired by applying an image processing method;

[0032] - a determination of the position of the terrestrial reference point in a frame of reference originating from a point in the acquired image; and,

[0033] - a determination of a geographical position of the detected terrestrial reference point by consulting a database associating a plurality of terrestrial reference points with their geographical position.

[0034] The position of the terrestrial reference point in a frame of reference originating from a point in the image and the geographical position of this terrestrial reference point define a line of sight between this terrestrial reference point and the vehicle (or more precisely between this terrestrial reference point and an image acquisition device on board this vehicle).

[0035] In the context of the invention, a "terrestrial reference point" is a point on Earth whose position (e.g., geographic coordinates) is known. In particular embodiments, the geographic position of this terrestrial reference point is an absolute position. In the field of navigation, a terrestrial reference point is also referred to as a "landmark".

[0036] This particular method of implementation makes it possible to overcome the drift of inertial navigation from positioning data from images acquired by the vehicle.

[0037] In certain implementation modes, the geographic position of the database's reference points is an absolute position. This characteristic is advantageous because it ultimately allows for the determination of an absolute, rather than relative, location of the vehicle. Furthermore, by using an absolute position, the orientation of the track does not need to be known.

[0038] In particular implementation modes, the tight hybridization unit is configured to estimate errors within the navigation data determined by the inertial navigation unit, this estimation being implemented from an error propagation model and from observations from the positioning data of the terrestrial reference point.

[0039] In specific implementation methods, the process further includes:

[0040] - an estimation, by the inertial navigation unit, of an inertial navigation state including said estimated navigation data, by application of a non-linear time evolution model;

[0041] - an estimate, by the tight hybridization unit, of a differential of the error on the inertial navigation state; and,

[0042] - a correction, by the inertial navigation unit, of the estimated state, as a function of the differential of the estimated error.

[0043] This error estimation involves determining a linearized time evolution of the differential of a state error and determining a differential of an innovation as a function of the state error. In other words, the linearized time evolution and the differential of the innovation as a function of the state error correspond to two models for estimating this error. This characteristic highlights the separation between inertial navigation, which calculates navigation data (speed, position, attitude, etc.), and the tight hybridization unit, which estimates the errors made by inertial navigation. To do this, a linearized Kalman filter is implemented, for example, to estimate the error state. This linearization is achieved by using the error differential. This error estimation is performed using measurements from the hybridization sensors (baro-altimeter, vision, GNSS).The error state thus estimated by the hybridization unit is then used to correct the inertial navigation. A corrected state is obtained. This feature is advantageous because it allows you to choose which error states to correct. For example, with vision, you might want to correct velocity and position errors, but not attitude errors. This feature also makes it easy to manage frequency differences between inertial navigation and hybridization sensor observations. Furthermore, this feature allows you to choose, when correcting inertial navigation using the error state estimated by the hybridization unit, between: - an "open loop" correction where the corrected state is only used for user output, i.e. the corrected state is not introduced into the inertial navigation, and - a "closed loop" correction where the corrected state is directly used in the inertial navigation. The "open loop" approach is generally chosen when hybridizing with sensors that are not very integrated, such as vision systems. This avoids any "contamination" of the inertial navigation system by the hybrid sensors and allows the system to take advantage of the accuracy of the navigation-grade inertial measurement unit.

[0044] According to a second aspect, the invention relates to a computer program comprising instructions for implementing a navigation method, when said program is executed by a processor.

[0045] This program can use any programming language, and be in the form of source code, object code, or code somewhere between source code and object code, such as in a partially compiled form, or in any other desirable form.

[0046] According to a third aspect, the invention relates to a computer-readable recording medium on which the computer program according to the invention is recorded.

[0047] The information or recording medium can be any entity or device capable of storing the program. For example, the medium can include a storage means, such as a ROM, for example a CD-ROM or a microelectronic circuit ROM, or a magnetic recording means, for example a hard drive.

[0048] On the other hand, the information or recording medium can be a transmissible medium such as an electrical or optical signal, which can be transmitted via an electrical or optical cable, by radio, or by other means. The program according to the invention can, in particular, be uploaded to a network such as the Internet.

[0049] Alternatively, the information or recording medium may be an integrated circuit (e.g., an FPGA, acronym for "Field-Programmable Gate Array") in which the program is incorporated, the circuit being adapted to execute or to be used in the execution of the process in question.

[0050] According to a third aspect, the invention relates to a navigation device configured to implement a method according to the invention.

[0051] The proposed navigation system has the advantages described above in relation to the proposed navigation method.

[0052] According to a fourth aspect, the invention relates to a navigation device comprising:

[0053] - a module for obtaining inertial data generated by a high-precision inertial measurement unit;

[0054] - an estimation module, by an inertial navigation unit of the navigation device, of navigation data from the inertial data obtained;

[0055] - a receiving module, via a tight hybridization unit of the navigation device connected to the inertial navigation unit, for positioning data of a terrestrial reference point within an image acquired by the vehicle, said positioning data being previously calculated by an image processing unit of said vehicle; and,

[0056] - a navigation data correction module, by applying a fusion, by the tight hybridization unit, of navigation data and positioning data of the terrestrial reference point.

[0057] According to a fifth aspect, the invention relates to a navigation system for a vehicle, said system comprising:

[0058] - a navigation device according to the invention;

[0059] - an inertial unit of measurement;

[0060] - an image acquisition device; and,

[0061] - a baro-altimeter.

[0062] The proposed navigation system has the advantages described above in relation to the proposed navigation method.

[0063] As mentioned previously, the image processing unit is configured to determine at least one position of a reference terrestrial point from the images acquired by the image acquisition device.

[0064] In specific implementation modes, the navigation system further includes a guidance module configured to guide the vehicle using corrected navigation data.

[0065] According to a sixth aspect, the invention relates to a vehicle comprising a navigation system according to the invention. Brief description of the drawings

[0066] Other features and advantages of the present invention will become apparent from the description below, with reference to the accompanying drawings which illustrate an example of an embodiment without being limiting in any way. In the figures:

[0067] [Fig.1] Figure 1 is a functional representation of a navigation system, according to an example of implementation of the invention;

[0068] [Fig.2] Figure 2 is a functional representation of the navigation system of Figure 1, according to an example of implementation of the invention;

[0069] [Fig.3] Figure 3 represents modules embedded in a navigation device of the navigation system of Figure 1, according to a particular mode of implementation of the invention;

[0070] [Fig.4] Figure 4 schematically represents an example of the hardware architecture of a navigation device from the navigation system in Figure 1;

[0071] [Fig. 5] Figure 5 represents, in flowchart form, a particular implementation method for a vehicle navigation system, for example, executed by the navigation device in Figure 3. Description of embodiments

[0072] The terms "first(s)", "second(s)", etc. are used in this document by arbitrary convention to enable identification and distinction of different elements considered in the embodiments described below, and do not imply any particular sequencing unless explicitly stated.

[0073] Figure 1 is a functional representation of a SYS navigation system, according to an example of implementation of the invention.

[0074] As illustrated in Figure 1, the SYS navigation system for a vehicle includes:

[0075] - a set of SENS sensors;

[0076] - a DISP navigation device configured to determine, from data provided by the sensors of the SENS assembly, CNAV vehicle navigation data. The DISP navigation device comprises an inertial navigation unit U_NAV connected to a tight hybridization unit U_HYB; and

[0077] - a CMD guidance module configured to guide the vehicle from the CNAV navigation data provided by the DISP device.

[0078] In a particular implementation, the SYS system is embedded in a vehicle, for example, a land vehicle (e.g., a car, a truck, a train), a boat, or an aircraft (e.g., an airplane, a helicopter, a plane, a drone). Specifically, in a preferred implementation, the SYS system is embedded in an aircraft.

[0079] As mentioned previously, in the context of the invention, the "navigation data" of a vehicle refers to data relating to the vehicle's position and / or movement and includes, for example, geographic coordinates (e.g., latitude, longitude, altitude), speed, and attitude. Navigation data can be defined absolutely with respect to the Earth's frame of reference, or relatively with respect to a reference position (e.g., a runway). By way of example, a relative position of the vehicle at a given time determined by the DISP navigation device may include one or more coordinates from the following set: an azimuth, a vertical distance, a longitudinal distance, and a lateral distance defined with respect to a reference position.

[0080] As illustrated in Figure 1, the SENS sensor set includes the following sensors:

[0081] - a high-precision inertial measurement unit IMU_HP;

[0082] - a BARO altimeter; and,

[0083] - an image processing unit U_IMG.

[0084] The high-precision inertial measurement unit (IMU_HP) provides inertial data to the inertial navigation unit (INV) U_NAV. In one embodiment, this inertial data comprises, for a plurality of measurement times, FS data relating to the specific force (IS, the sum of external forces other than gravitational forces divided by the mass) and the angular velocity IL of the vehicle. Typically, the IMU_HP high-precision inertial measurement unit includes: three gyroscopes measuring the three components of the angular velocity IL (rates of change of roll, pitch, and yaw angles); and three accelerometers measuring the three components of the specific force.

[0085] In a specific implementation, this inertial measurement unit is of "navigation" grade. As mentioned previously, a navigation-grade inertial measurement unit differs from a standard inertial measurement unit in its accuracy, stability, and reliability. Indeed, a navigation-grade inertial measurement unit typically exhibits a gyroscopic drift class of 10 -2 degrees per hour, and a position error drift class of 1 nautical mile per hour. In contrast, standard inertial measurement units typically exhibit a gyroscopic drift class of a few degrees per hour, and a position error drift class of a few tens of nautical miles per hour.

[0086] The BARO altimeter outputs altimetric data. In one embodiment, the BARO altimeter is a barometric altimeter. Specifically, the altimetric data represents the vehicle's altitude or changes in its altitude over a range of measurement times. Alternatively, the altimeter measures atmospheric pressure, which, combined with an altitude / pressure profile, provides data relating to the vehicle's altitude.

[0087] The UJMG image processing unit outputs VIS positioning data. In a specific implementation, the UJMG image processing unit includes, or is configured to communicate with, a BDD storage medium and a CAM image acquisition device. The BDD storage medium, for example, a database, contains the positions (Legion, geographic coordinates) of a plurality of ground reference points (hereafter referred to as "landmarks") as well as information relating to graphical representations of the ground reference points. For example, ground reference points could be points on a runway, a navigation light, a Precision Approach Path Indicator, etc.The CAM image acquisition system includes at least one camera mounted in the vehicle and equipped with an electromagnetic radiation sensor whose wavelengths belong to the visible light and / or infrared spectrum. In one embodiment, the CAM image acquisition system includes at least one of the following cameras: a visible light camera; a near-infrared camera; a short-wavelength infrared camera; a medium-wavelength infrared camera; or a long-wavelength infrared camera. The CAM image acquisition system is configured to acquire a plurality of images for a plurality of measurement times.

[0088] The U_IMG image processing unit takes as input images acquired by the CAM image acquisition device and LOC_VIS navigation data from the BDD recording medium. More precisely, in a particular implementation mode, the U_IMG image processing unit is configured to:

[0089] - obtain an image acquired by the CAM image acquisition device;

[0090] - detect a reference point within this image, by applying an image processing method;

[0091] - determine the position of the terrestrial reference point in a frame of reference originating from a point in the acquired image, for example in pixel coordinates; and,

[0092] - determine a geographical position of the detected terrestrial reference point by consulting the BDD database associating a plurality of terrestrial reference points with their geographical position.

[0093] Thus, the position of the terrestrial reference point in a frame of reference originating from a point in the image and the geographical position of this terrestrial reference point define a line of sight between this terrestrial reference point and the vehicle (or more precisely between this terrestrial reference point and an image acquisition device on board this vehicle).

[0094] As illustrated in Figure 1, in one embodiment, the DISP navigation system comprises an inertial navigation unit U_NAV and a tight hybridization unit U_HYB, for example, in the form of an error Kalman filter. The inertial navigation unit U_NAV is configured to integrate in time the specific force and angular velocity data FS produced by the inertial measurement unit IMU_HP, and thus estimate navigation data, for example, in terms of vehicle position (POS), velocity (VEL), and attitude (ATT). In a particular implementation, the inertial navigation unit U_NAV is configured to determine an absolute 6D pose (3D position and 3D attitude) and absolute velocity.

[0095] The tight hybridization unit U_HYB determines, from these estimated navigation data and from the VIS positioning data from the image processing unit UJMG, corrections δPos, δVel, to be applied to the position data POS, and velocity data VEL.

[0096] In a specific implementation mode, the CNAV navigation data provided as output by the DISP navigation device corresponds to:

[0097] - to the ATT navigation data in terms of attitude estimated by the inertial navigation unit U_NAV; and

[0098] - to the navigation data in terms of POS position and VEL velocity estimated by the inertial navigation unit U_NAV, to which corrections are applied, determined by the tight hybridization unit U_HYBRID based on the ALT altitude data provided by the BARO barometer and the VIS positioning data provided by the image processing unit. In other words, the tight hybridization unit U_HYBRID compensates for inertial navigation drift (i.e., recalibrates) using data from the CAM and BARO sensors.

[0099] In other words, the ATT navigation data in terms of attitude provided to the vehicle guidance module corresponds to that initially determined by the inertial navigation unit U_NAV, without a correction of this attitude being determined by the tight hybridization unit.

[0100] It is important to note that the navigation data determined at any given time is dependent on the navigation data determined at previous times. This is because the inertial navigation unit (U_NAV) integrates over time the specific force and angular velocity (FS) data produced by the high-precision inertial measurement unit (IMU_HP) to determine CNAV navigation data. Therefore, small measurement errors of the specific force and angular velocity are integrated over time by the inertial navigation unit, leading to increasing velocity and position errors (i.e., inertial navigation drift). Consequently, using data from the image processing unit (U_IMG) at a given time to recalibrate the high-precision inertial measurement unit (UJMG) improves the accuracy of the navigation data determined at subsequent times.

[0101] It should also be noted that in a particular implementation mode, the SYS navigation system includes or is configured to communicate with a GNSS satellite positioning system (not shown) configured to output positioning data. In this specific case, the SYS navigation system is then configured to activate the hybrid inertial-vision navigation solution only when the SYS navigation system is unable to use the data from the GNSS satellite positioning system to compensate for drift (i.e., recalibrate) of the high-precision inertial measurement unit U_IMG.

[0102] Finally, it should be noted that other functional representations and / or architectures can be considered. Indeed, in a particular implementation mode, all or part of the sensors of the SENS system and / or the CMD guidance module are integrated within the DISP navigation system. Thus, the U_IMG image processing unit, for example, is integrated within the DISP navigation system.

[0103] Figure 2 is a functional representation of the SYS navigation system of Figure 1, according to an example of an implementation of the invention. This figure is described in more detail below.

[0104] Figure 3 represents modules embedded in a DISP navigation device of the SYS navigation system of Figure 1, according to a particular embodiment of the invention.

[0105] As illustrated in Figure 3, the DISP navigation device of the SYS navigation system includes, in particular, an inertial navigation unit U_NAV including:

[0106] - a MOD_OB_IN module for obtaining inertial data generated by a high-precision inertial measurement unit;

[0107] - a MOD_EST_NAV module for estimating, by an inertial navigation unit of the navigation device, navigation data from the inertial data obtained.

[0108] The DISP navigation system also includes a tight hybridization unit U_HYB connected to the inertial navigation unit U_NAV and including:

[0109] - a MOD_OBJMG module for receiving positioning data from a terrestrial reference point within an image acquired by the vehicle, said positioning data being previously calculated by an image processing unit of said vehicle; and,

[0110] - a MOD_DET_CNAV module for correcting navigation data, by applying a fusion of navigation data and positioning data from the terrestrial reference point.

[0111] Their functionalities are described in more detail below with reference to different implementation methods.

[0112] Figure 4 schematically represents an example of the hardware architecture of a DISP navigation device of the SYS navigation system.

[0113] As illustrated in Figure 4, the DISP navigation device has the hardware architecture of a computer. Thus, the DISP navigation device includes, in particular, a processor 1, random access memory 2, read-only memory 3, and non-volatile memory 4. It also has communication means 5.

[0114] The read-only memory 3 of the DISP navigation device constitutes a storage medium according to the invention, readable by the processor 1, on which a computer program PROG according to the invention is stored, comprising instructions for executing steps of the navigation process according to the invention. The PROG program defines functional modules of the DISP navigation device, which rely on or control the hardware elements 1 to 5 of the DISP navigation device mentioned above. These functional modules are illustrated in Figure 3 by way of no limitation and are described in more detail below with reference to different implementation methods.

[0115] In the implementation modes described below, the communication means 5 enable the DISP navigation device to obtain inertial data FS, altimetric data ALT, and positioning data VIS from the SENS sensor array. The communication means 5 also enable the DISP navigation device to transmit CNAV navigation data to the CMD guidance module. To this end, the communication means 5 include a wired or wireless communication interface capable of implementing any suitable communication protocol.

[0116] Figure 5 represents, in the form of a flowchart, a particular method of implementing a vehicle navigation process, for example executed by the DISP navigation device of Figure 3.

[0117] Notations

[0118] x: exact (or "true") value;

[0119] x: estimated value;

[0120] x: measured value;

[0121] sx: composition between an exact value and an estimated value;

[0122] dx: differential of x;

[0123] x a : projection of x in a coordinate system [a];

[0124] [t]: terrestrial reference frame;

[0125] [5]: local geographical reference;

[0126] [c]: CAM image capture device reference;

[0127] [m]: reference frame of the inertial measurement unit;

[0128] [r]: local reference frame attached to the landing strip;

[0129] [i]: a Galilean frame of reference fixed relative to the stars, in which the laws of Newtonian mechanics apply. The origin and axes of the inertial frame [i] are arbitrary. In the following description, we consider the special case where its origin is the center of mass of the Earth, its Z-axis is collinear with the Earth's axis of rotation, its X-axis is located in the equatorial plane and oriented towards the vernal equinox, and its Y-axis completes the right-handed orthonormal trihedron.

[0130] To: latitude;

[0131] φ: longitude;

[0132] h: altitude;

[0133] R n : radius of curvature North;

[0134] R e : radius of curvature East;

[0135] O: center of the Earth;

[0136] M: navigation point;

[0137] r = OM: vehicle position vector; dv

[0138] v = dr / dt (with respect to [t]): velocity vector with respect to the Earth;

[0139] g(r): gravity vector;

[0140] f m : specific strength;

[0141] b a : accelerometer bias;

[0142] d g : gyroscopic drift;

[0143] T ab : matrix of the SO(3) group ( a = T ab ■ x b );

[0144] φ: rotation vector, associated with a SO(3) rotation matrix by T = expm(A(φ)) 0 —zy '

[0145] A([x,y,z] T ) = z 0 —x antisymmetric matrix ("skew-symmetric matrix") -yx 0 (according to Anglo-Saxon terminology);

[0146] I n : Identity matrix of dimension n;

[0147] O n : Null matrix of dimension n;

[0148] (has a J b: rotation vector of the frame [a] with respect to [b] projected onto [c].

[0149] The navigation process includes a first step S100 during which inertial data FS, Ω generated by a high-precision inertial measurement unit are obtained (e.g., received) by the navigation device DISP. This step is implemented, for example, by the MOD_OB_IN module for obtaining inertial data from the inertial navigation unit U_NAV of the navigation device. Subsequently, these measured data are denoted ũ.

[0150] The navigation process further includes a step S110, implemented by the inertial navigation unit U_NAV, during which the state x IRSThe inertial navigation state of the system (also called the "nominal state") is determined. More precisely, this step is implemented, for example, by the MOD_EST_NAV module of the DISP navigation device. To do this, a non-linear time evolution x IRS of state x IRS The inertial navigation path is determined. More formally, this temporal evolution x can be expressed, for example, as follows: 1RS = f COIRS' )

[0151] This formalization of the temporal evolution x is also represented in figure 2.

[0152] The state of inertial navigation x IRS can be expressed, for example, in the form of a vector comprising:

[0153] - positional data including a matrix T gt of group SO(3) and an altitude h, with T gt = f(rg), r g the vehicle position vector in a local reference frame [5], h E IR;

[0154] - a speed v g such that v g E 3 ;

[0155] - a T attitude gm , with T gm a matrix of SO(3);

[0156] - a bias b a of the accelerometer such that b a E 3 ; And

[0157] - a gyroscopic drift such that d g E 3 ;

[0158] The nonlinear time evolution equations of the inertial navigation state can then be expressed, for example, as follows:

[0159] - temporal evolution of position data

[0160] - temporal evolution of altitude = v g •

[0001] T

[0161] - temporal evolution of speed v g = Tgm ' (fm + b a ^ + g g (?g) ~A Ç(V g +

[0162] - temporal evolution of attitude 7* _ 7* j ( I Çj _ I 1 gm 1 gm ' 71 L ' u gb

[0163] - temporal evolution of the accelerometer bias ia = 0

[0164] - temporal evolution of gyroscopic drift â5= o

[0165] with A being the antisymmetric matrix function; a) g ^ the rotation vector of the local frame [g] with respect to the terrestrial frame [t] projected onto the local frame [g]; f m the specific force measured by the inertial measurement unit; b a accelerometer bias; g g (rg) the gravity vector; r g the position vector projected onto the vehicle's frame of reference [g]; a> g l the rotation vector of the terrestrial frame of reference [t] with respect to the frame of reference [i] projected into the local frame of reference [g]; the angular velocity measured by the inertial measurement unit; the rotation vector of the terrestrial frame [t] with respect to the local frame [g] projected into the frame [m] of the inertial measurement unit;

[0166] ω̂g t / i = T̂ gt · [Ω 0 0] T the angular velocity of the Earth (Ω being known);î 0 R î e +h 0

[0167] M 9 / t = M pg ■ v g with M pg = R n +h 0 0 and  latitude, 0 longitude, R n the tan 0 0 Re+h radius of curvature North and R e the radius of curvature East,

[0168] g g {fg) = f(r g ) And, [ 0169] -f g T m - M pg - v g -

[0170] Returning to Figure 5, the process further includes a step S120 during which the CAM image capture device generates an image stream. The images in this stream are processed on the fly by the UJMG image processing unit (steps S130 to S150 described below). More precisely, during step S130, a landmark is detected in one of the images in this stream. This detection is implemented, for example, by applying an image processing method, such as the Scale-Invariant Feature Transform (SIFT); the Speeded Up Robust Features (SURF); the Hough transform; or other object detection methods within an image, for example, those based on deep learning principles.

[0171] During step S140, the position of the landmark is determined within a reference frame originating from a point in the acquired image. This position is expressed, for example, in "pixel coordinates." Then, during step S150, a geographic position of the landmark detected during step S130 is determined. To do this, a database (BDD) associating a plurality of terrestrial reference points with their geographic positions is consulted.

[0172] In a particular embodiment, the database provides an absolute position of a landmark, that is, in an absolute reference frame, such as the Earth's reference frame. This condition allows the invention to determine an absolute, rather than relative, location of the vehicle. Furthermore, by using an absolute position, the runway orientation does not need to be known.

[0173] The position of the landmark in a reference frame originating from a point in the image, and the geographical position of this landmark, for example in relation to the terrestrial reference frame, define a line of sight between this landmark and the vehicle.

[0174] In a specific implementation, navigation data (position, speed, attitude) is used to aid in landmark detection, specifically in the S130 landmark detection step. In particular, image processing can use this data to narrow the search area for reference points within the image.

[0175] The process further includes a step S160 in which the positioning data for a landmark are obtained (e.g., received or accessed) by the tight hybridization unit U_HYB of the DISP navigation device. This obtaining (e.g., receiving or accessing) step S160 is, for example, implemented by the MOD_OB_IMG module of the tight hybridization unit U_HYB of the DISP navigation device.

[0176] As mentioned previously, the inertial navigation unit U_NAV therefore estimates an inertial navigation state x IRS Now, this estimated state x IRSThis involves errors. The tight hybridization unit U_HYB therefore relies on the navigation marker positioning data obtained from image processing to estimate these errors using an error propagation model, and to determine corrections to apply to the navigation data previously estimated by the inertial navigation unit. As mentioned earlier, this tight hybridization unit can take the form of an error Kalman filter. Alternatively, an extended Kalman filter, an invariant Kalman filter, particle filtering, the least squares method, or a PID (Proportional, Integral, Derivative) corrector can be considered.

[0177] The error sr g In terms of position, this can be expressed, for example, as follows:

[0178] sr g = T gt ■ [r t — r t ], and its differential is expressed as dsr g = T gt ■ Dr. t .

[0179] The sv error g In terms of position, it is expressed, for example, as follows: sv g = v g — v g , and its differential is expressed as dsv g = dv g .

[0180] The error in terms of attitude is expressed, for example, as an error rotation vector. £ <p™ / 3 e R 3 , and its differential is expressed such that (p g = of^^ 3 .

[0181] The sb error a In terms of accelerometer bias, it is expressed, for example, as sb a = b a — b a , and its differential is expressed as: dsb a = db a .

[0182] The SD error g In terms of gyroscopic drift, it is expressed, for example, as sd g = d g — d g , and its differential is expressed as dsd g = dd g .

[0183] Thus, the state of the inertial navigation error, denoted dsx, which the tight hybridization unit U-HYB must estimate is therefore given by: d£r g d.£V i n J

[0184] d£X = < Pg deb a d£dg

[0185] The process includes a step S170 for determining the temporal evolution of the inertial navigation error by applying a propagation model FQ. This evolution d£X is expressed, for example, as follows: d£X = F(X IRS ). d£X

[0186] This formalization of the temporal evolution of the inertial navigation error d£X is also represented in figure 2.

[0187] Since the time evolution of errors is non-linear, the equations must be linearized.

[0188] The temporal evolution of g Positional error can be expressed, for example, as follows: d£r n = d£Vn - Ll ( • d£r n î 0 R e +h

[0189] withM w = 0 0 tan 0 R e +h

[0190] The temporal evolution of V g The speed error is expressed, for example, as follows: d£Vg = -A < T gm ■ (f m + ' (Pg d" Tgm ' + [-^(^5) ' Mpg ~ ^( M g + • d£Vg + [Mgg + 2 A(V g) ' AÇM^ ' M gg] ■ d£r g

[0191] with M pg = M gg

[0192] The temporal evolution (p g An error in attitude can be expressed, for example, as follows: (Pg = T gm .ddg -A (^g / l ) ■ < Pg - Mpg ' d£Vg ~ A ' Mgg '

[0193] And the temporal evolution of the error resulting from accelerometer bias and gyroscopic drift is expressed, for example, as follows: dd r , = db n = 0

[0194] In this case, the matrix F is expressed as follows: / £\

[0195] F(dEr g ,d£rg) =

[0196] F(dEr g ,dEVg) = I3

[0197] F(dzv g , derj = [M gg + 2 A(y g ) • • M gg ]

[0198] F(dsv g , dEV g ) = [71(vJ • M pg - A(a)f + 2&$)]

[0199] F(dev g , (pg) = —A (î gm ■ (J™ + S a )}

[0200] F(dEV g i deb a ) = T gm ^gj ' M gg

[0202] FÇ(p g ,dEV g ) = -M pg

[0203] F( <p g , <p g ) = -A^f)

[0204] F (pg ,dEd g ) = T gm

[0205] F(dEb a :) = 0

[0206] F(dEd g :) = 0

[0207] The navigation process further includes an S180 step during which the tight hybridization unit uses the landmark positioning data to correct the navigation data previously estimated by the inertial navigation unit. An observation model is therefore defined, which can be expressed, for example, in the form y = h( ).

[0208] Innovation is defined, for example, as the difference between a measured observation and an observation reconstructed by estimating the inertial navigation unit. In a particular implementation mode, during this S180 step, a differential of the innovation dey is calculated as a function of the inertial navigation error on the inertial navigation state. More formally, the differential of the innovation dey is expressed, for example, as: dey = H(X IRS ). dex

[0209] This formalization of the innovation differential dey is also represented in figure 2.

[0210] It is important to note at this stage that the navigation method according to the invention can be implemented as soon as a single visual landmark is captured by the image capture device. In one particular embodiment, this landmark corresponds to a corner of the runway. Alternatively, this landmark corresponds to a virtual point, such as the center of gravity of the runway. Generally, this landmark corresponds to a point attached to the runway, measurable, and whose geographical position, for example relative to the Earth's reference frame, is known and provided by the database.

[0211] In the following description, the landmark in question corresponds to the center of gravity of the runway. In this particular implementation, the measured observation corresponds to the coordinates of a pixel. It is denoted y. ba However, it should be noted that there are no limitations attached to the type of measured observation considered. Furthermore, the geographical position rb ar The position of the landmark in the Earth's frame of reference is assumed to be known, for example after consulting the BDD database. The observation equation is then expressed as y bar = h bar x)'.

[0212] In the following description, the CAM image capture device is modeled as a pinhole camera. It should be noted, however, that there are no limitations attached to the model of the image capture device considered. In this particular case, m x ex ^cz

[0213] y b „ = [“] = [“'] + / ■ h C y m y ez -

[0214] with u, v the two coordinates of a pixel representing the position of the landmark within the image;

[0215] u0, v0 the coordinates of the center of the image plane;

[0216] f is the focal length;

[0217] m x,m y intrinsic parameters of the CAM image capture device;

[0218] The line of sight h c The relationship between Tamer and the CAM image capture device, projected into the CAM image capture device's frame of reference, can then be expressed, for example, as follows:

[0219] h c = T cm T mg T gt • [r bar - rf] - T cm • L m

[0220] with the position vector of the carrier; and L m the lever arm between the CAM image capture device and the IMU_HP inertial measurement unit, whose error is assumed to be negligible.

[0221] The following measurement model is also considered: u = û + bbu

[0222] v = v + bbv

[0223] with u, vies exact (or "true") values; ü, vies measured values; etbbu, bbvies measurement bias.

[0224] The innovation sy, which corresponds to a difference between a measured observation and an observation reconstructed by estimation, is expressed, for example, as: sy = ÿ bar — -bar (. -IRs)-

[0225] The innovation differential can be expressed, for example, as follows: du0] \dbb,,

[0226] dsy bar = dv '.o.,dbb v ^uvhc ' ^cm ' h m L m ^ ■ % ^uvhc ' ^cm ' ^mg hg^) ' < Pg 4 M UV h C T 1 cm f 1 mg ' [•^( T gt ' / g] • dETg + M uvflc 'T cm' T m g fj r bar u 'g

[0227] where the state duv is composed of du0 and dv0; the state dbb uv is composed of dbb u and dbb v Mr. uvflc= f(hc); represents the angular harmonization error between the inertial measurement unit IMU_HP and the CAM image capture device; dTg ar represents the position error of the landmark provided by the BDD database.

[0228] The error state vector can then be expressed as follows: der g dev g < Pg beginning a

[0229] dsx = dedg lev dbb uv fj r bar L“fg

[0230] It is important to note here that components of duv, dbb uv , and dTg ar have therefore been added to the error state vector previously mentioned in reference to step S170. As mentioned below, the addition of components to the error vector dsx entails the addition of a propagation model for these components, and the error state propagation matrix F is therefore also modified.

[0231] Furthermore, the observation matrix H is, for example, formalized by the following equations:

[0232] H(:,d£r g ) = M uvhc 'Tcm' T m g [4( Tg t • H t ) • M ng -Z3]

[0233] H(:,d£Vg) = 03 [ 0234] H(J., (Pg') — ^uvhc ' T cm ' T m g ' hg)

[0235] W(:,d£h a ) = 03

[0236] H.,d£dg = 03

[0237] H(:,duv) = / 3

[0238] H,dbb uv ) = ~ I3

[0239] / / (:, ) = ~M uvhc - T cm - A h m - L m )

[0240] H(:, dr g bar ) = M uvflc • T cm • T mg

[0241] The navigation process finally includes a so-called "looping back" step S190 during which the dsx corrections determined during step S180 are applied, in order to correct the x state IRS of the inertial navigation initially estimated by the inertial navigation unit U_NAV.

[0242] This S190 step can be expressed, for example, as follows: {x IRS ,dsx} «- b(x IRS , d£x). This formalization of the looping step is also represented in Figure 2.

[0243] The looping function bQ can be expressed, for example, through the following looping equations for position, altitude, and velocity:

[0244] - T position re-looping gt <- expm^A(M rig • d£r g ^ T gt withexpmQ the function exponential of a matrix;

[0245] - altitude loop h «- h. + dsh, with dsh = d£r g •

[0001]

[0246] - speed re-looping v g “- v g + d£V g

[0247] In a particular implementation, the bQ function does not include a feedback equation related to attitude, accelerometer bias, and gyroscopic drift. This characteristic is linked to the use of a high-precision inertial measurement unit, which eliminates the need for these feedback equations.

[0248] Alternatively, the loopback function bQ further includes:

[0249] - attitude feedback: T gm «- expm (jiÇ <p g ^ ■ T gm

[0250] - feedback of accelerometer biases: b a “- b a + deb a

[0251] - Gyroscopic drift feedback: d g “- d g + d£dg

[0252] As mentioned previously, the components were added to the error state vector during step S180.

[0253] As mentioned below, the addition of duv and dbb components uv , and dTg ar The addition of a propagation model for these components to the error vector dsx necessitates the modification of the error state propagation matrix F. More precisely, a time evolution model is established for these new components. The following constraints can be considered, for example:

[0254] - duv can be considered constant, therefore F(:, duv) =

[0255] - dbb uv can be modeled by white noise, therefore F(:, dbb uv ) = O3

[0256] - can be considered constant, therefore F(:, ) = / 3

[0257] - drg arcan be considered constant, therefore F(:, drg ar ) = / 3

[0258] Results

[0259] Since only a landmark is required to implement this navigation procedure, it can be initiated even when the aircraft is a significant distance from the runway. More precisely, the final approach segment from the Final Approach Fix (FAF) to the Decision Altitude (DA) takes an average of 4 minutes for a distance of 7 NM (~13 km) for a typical Airbus A320 airliner. Some prior art methods only provide system availability from 2 km from the runway, or from 855 meters above the DA at 200 feet for a typical 3° descent path. Considering the average speed of an approaching airliner—115 knots over this FAF / DA segment—the system is only available during the last 15 seconds before reaching the DA. The percentage of availability from the FAF point to the DA decision altitude is then 6.25%.

[0260] Under the same conditions, the invention offers availability starting from 6 km from the runway, or from 4.855 km above the decision altitude DA (~2.6 NM). This corresponds to an availability of 82 seconds for the final segment. The availability percentage is then 34.2%.

[0261] The invention has so far been described in the case where a single landmark is captured by the CAM image capture device and detected by the UJMG image processing unit. However, the invention remains applicable in the case where several landmarks (for example, several corners of the runway) are captured and detected.

[0262] The invention has also been described so far in cases where global data—such as position or attitude—determined relative to an absolute reference frame are used. However, the invention remains applicable when local data, e.g., determined relative to a relative reference frame, are used. Indeed, the transition from a global reference frame to a local reference frame is straightforward, for example, from knowledge of the position and orientation of the runway.

Claims

Demands

1. A navigation method for a vehicle, the method being implemented by a navigation device (DISP) and comprising: - obtaining (S100) inertial data generated by an inertial measurement unit of navigation grade (IMU_HP); - an estimation (S110), by an inertial navigation unit (U_NAV) of the navigation device (DISP), of navigation data from the inertial data obtained; - a reception (S160), by a tight hybridization unit (U_HYB) of the navigation device (DISP) connected to the inertial navigation unit (U_NAV), of positioning data of a single terrestrial reference point within an image acquired by the vehicle, said positioning data being previously calculated by an image processing unit of said vehicle determining an absolute geographic position of the single terrestrial reference point; and, - a correction (S190) of the navigation data, by applying a fusion, by the tight hybridization unit (U_HYB), of the navigation data and the positioning data of the only terrestrial reference point.

2. A navigation method according to claim 1, the vehicle further comprising a global navigation satellite system (GNSS), the method being implemented when said global navigation satellite system is unusable.

3. A navigation method according to claim 1 or 2, wherein the navigation data determined by the inertial measurement unit of navigation grade are not all corrected.

4. A navigation method according to any one of claims 1 to 3, wherein the estimated navigation data includes a vehicle attitude that is directly exploited by a vehicle guidance module (CMD), without being corrected.

5. A navigation method according to any one of claims 1 to 4, further comprising the following steps, implemented by the image processing unit (UJMG): - an acquisition (S120) of the image by the vehicle;- a detection (S130) of the terrestrial reference point (AMER) in the acquired image by application of an image processing method; - a determination (S140) of a position of the terrestrial reference point (AMER) in a reference frame originating from a point in the acquired image; and, - a determination (S150) of a geographical position of the terrestrial reference point (AMER) detected by consulting a database associating a plurality of terrestrial reference points with their geographical position.

6. A navigation method according to claim 5, wherein the geographic position of the terrestrial reference points in the database is an absolute position.

7. A navigation method according to any one of claims 1 to 6, wherein the tight hybridization unit (U_HYB) is configured to estimate errors within the navigation data determined by the inertial navigation unit (U_NAV) from an error propagation model (F) and from observations (y) from the positioning data of the single terrestrial reference point.

8. A navigation method according to any one of claims 1 to 7, further comprising: - an estimate (S110), by the inertial navigation unit (U_NAV), of an inertial navigation state including said estimated navigation data, by application of a non-linear time evolution model; - an estimate (S170, S180), by the tight hybridization unit, of a differential of the error on the inertial navigation state; and, - a correction (S190), by the inertial navigation unit (U_NAV), of the estimated state, as a function of the differential of the estimated error.

9. Computer program (PROG) comprising instructions for implementing a navigation method according to any one of claims 1 to 8, when said program is executed by a processor.

10. Computer-readable recording medium on which a computer program according to claim 9 is recorded.

11. Navigation device (DISP) configured to implement a method according to any one of claims 1 to 8.

12. Navigation system (SYS) for a vehicle, said system comprising: - a navigation device (DISP) according to claim 11; - an inertial measurement unit (IMU); - an image acquisition device (CAM); and, - a baro-altimeter (BARO).

13. Vehicle comprising a navigation system (SYS) according to claim 12.

Images