Method for calibrating a gyrometer of an object, object, computer program product and storage means - Patents.com

Abstract

The present invention relates to a method for calibrating a gyrometer (11) of an object (1) moving in an ambient magnetic field, the method comprising: (a) obtaining a measurement of angular velocity by a gyrometer (11) and at least two components of the magnetic field by magnetic measurement means (20) fixed to said object (1); (b) determining a value of at least one calibration parameter of the gyrometer (11) that minimizes a first equation defined by an estimated angular velocity of the object (1) and at least one first magnetic equation of magnetic field components, wherein the estimated angular velocity is a function of the measured angular velocity and the calibration parameter, and the at least one first magnetic equation assumes that the magnetic field is uniform and stationary.

Description

[Technical field]

[0001] The present invention relates to GNSS-less navigation.

[0002] More precisely, the invention relates to a method for calibrating a gyrometer fixed to a magnetometer. [Background technology]

[0003] Nowadays it is common to track the position of vehicles via GNSS (Global Navigation Satellite System, e.g. GPS) or using wireless communication networks (triangulation with radiating terminals, wifi networks, etc.).

[0004] These methods are very limited because they cannot guarantee the availability and accuracy of the information, one and the other being affected by any masking between the source and the receiver, and they rely on external technologies, such as satellites for GNSS, which may be unavailable or even spontaneously scrambled.

[0005] Alternatively, "autonomous" methods are also known that track the relative displacement of a vehicle in any environment, for example due to inertial or magnetic inertial units. By "relative displacement" is meant the trajectory of the vehicle in space relative to a point at initialization and a data marker. In addition to the trajectory, these methods also make it possible to obtain the orientation of the vehicle relative to the same initial marker.

[0006] A navigation class inertial unit is formed from at least three accelerometers and three gyrometers with a three-axis arrangement. Typically, the gyrometers "hold" the marker, and a double time integration of the accelerometer measurements makes it possible to estimate the motion.

[0007] It is known that it is necessary to use very precise sensors in order to be able to use conventional inertial navigation, especially as implemented in complex applications such as the navigation of fighter and civil aircraft, submarines, naval ships, etc. Indeed, the double time integration of acceleration measurements means that a constant error in acceleration produces an error in position that increases proportional to the square of the time.

[0008] Alternatively, if velocity vector information in the carrier's markers can be provided by an external source (vehicle odometer readings, ship logs, aircraft pitot tubes), it is known that it is possible to obtain the carrier's trajectory when it is possible to know the carrier's orientation, and in particular its heading.

[0009] In a certain number of cases, the initial orientation is known (initial "alignment" of the inertial device) and can be obtained by sensors other than inertial (e.g., magnetic) or can be derived directly from the inertial sensors if they allow for accurate measurement of ground rotation.

[0010] The difficulty, when it is desirable to work with low-cost sensors, is to “increase” their performance by continuously calibrating their error parameters in such a way that the orientation information of the carrier, and in particular the ground vehicle's orientation information, is maintained as good as possible.

[0011] Calibrating the gyrometer using magnetic measurements has therefore been proposed in this context.

[0012] US Patent No. 5,399,633 proposes checking whether the change in the magnetic field between two measurement points corresponds to the change in attitude predicted by the gyrometer integration. However, this assumes that the magnetic / accelerometer measurements are reliable, which is tricky. In practice, the following is observed: - The surrounding magnetic field is not always uniform and stationary; Nearby metal elements (e.g. bodywork) or magnets affect the magnetic measurements (soft and hard iron type effects). See also US Pat. No. 5,399,633. - To accurately calibrate the magnetometer, it is necessary to place it in the maximum spatial orientation so as to "excite the error model" in order to obtain observability. This problem is easily solved in the case of a device that can be carried by hand and turned in any direction (see, for example, patent document 3). If the magnetometer is mounted on a vehicle, it has the following limitations: Due to the fact that during a typical run of the vehicle, roll and pitch deviate slightly from zero. Due to the fact that the body is magnetized or the vehicle is loaded, this means that the calibration may no longer be appropriate; this requires regular recalibration.

[0013] US Pat. No. 5,399,433 is more relevant to smartphone-type devices with a magnetic compass and a gyrometer and addresses the problem of magnetic disturbances. This patent is rather aimed at the calibration of a magnetic compass using gyrometric data, but takes as a starting point the quaternion of the magnetic acceleration measurement and recalibrates the angular velocity ω=2q from this quaternion. -1 q · However, the assumption that the magnetometry / accelerometry measurements are reliable still holds, and in the opposite case we simply propose not to update the gyrometer bias or to use the default quaternion stored in memory as long as this situation persists.

[0014] It would be desirable to have a new method for calibrating the gyrometers of an object for the purpose of estimating the motion of the object, which allows for good quality of results and is not limiting. [Prior art documents] [Patent documents]

[0015] [Patent Document 1] US2012 / 0116716 [Patent Document 2] US2004 / 0123474 [Patent Document 3] FR1757082 [Patent Document 4] US2011 / 0178707 Summary of the Invention

[0016] Thus, the present invention relates according to a first aspect to a method for calibrating a gyrometer of an object moving in an ambient magnetic field, said method being characterized in that it comprises the following steps: (a) an obtaining step, - the measured angular velocity of the object by the gyrometer, and of at least two components of the magnetic field and / or of the i-th derivative of the magnetic field around a magnetic measuring means by a magnetic measuring means fixed to said object, The acquisition steps and; (b) determining, by the data processing means, values ​​of at least one calibration parameter of the gyrometer that minimize a first equation defined by the estimated angular velocity of the object and at least one first magnetic equation of a component of the magnetic field and / or an i-th order derivative of the magnetic field, the estimated angular velocity of the object is a function of the measured angular velocity and the calibration parameters of the gyrometer; - the at least one first magnetic equation assumes that the magnetic field is uniform and stationary around the magnetic measuring means; The determining step.

[0017] Other advantageous and non-limiting features include: Estimated angular velocity of the object ω (estimation) gyro is the model

number

number

number

number

number

number

number

number

number

number

[0018] According to a second aspect, an object moving in an ambient magnetic field, comprising a gyrometer configured to obtain a measured angular velocity of said object, and magnetic measurement means configured to obtain at least two components of the magnetic field and / or the i-th derivative of the magnetic field, said object comprising data processing means comprising: - data processing means configured to determine values ​​of at least one calibration parameter of the gyrometer that minimize a first equation defined by at least one first magnetic equation relating an estimated angular velocity of the object and a component of the magnetic field and / or an i-th order derivative of the magnetic field, The estimated angular velocity of the object is a function of the measured angular velocity and the gyrometer calibration parameters, o At least one first magnetic equation assumes that the magnetic field is uniform and stationary around the magnetic measuring means.

[0019] Other advantageous and non-limiting features include: The object further comprises additional acquisition means configured to acquire a measured linear velocity of the object, the data processing means further comprising: - Estimating parameters that represent the errors on the calibration parameters; - configured to redetermine, if said parameter representative of the error is greater than a predefined threshold, a value of at least one calibration parameter of the gyrometer that minimizes a second equation defined by at least one second magnetic equation relating the estimated angular velocity of the object and a component of the magnetic field and / or an i-th order derivative of the magnetic field, the at least one second magnetic equation assuming that the magnetic field is stationary around the magnetic measuring means; The object is a vehicle having wheels and the additional acquisition means are at least two odometers mounted on the wheels of the vehicle.

[0020] According to a third and a fourth aspect, it is proposed a computer program product comprising code instructions for the execution of the method according to the first aspect of the invention for calibrating a gyrometer; and storage means readable by a part of a computing device, the computer program product comprising code instructions for the execution of the method according to the first aspect of calibrating a gyrometer. [Brief description of the drawings]

[0021] Other characteristics and advantages of the invention will appear on reading the following description of preferred embodiments, which description is given below with reference to the accompanying drawings, in which:

[0022] [Figure 1] 2 shows an example of a vehicle architecture for the implementation of the method according to the invention; [Diagram 2] FIG. 2 illustrates the steps of a preferred embodiment of the method according to the invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0023] architecture

[0024] 1, the method allows the calibration of the gyrometer 11 of an object 1 moving in an ambient magnetic field, typically the Earth's magnetic field (modified by nearby metal objects, if applicable), described as M. As already explained, the magnetic field is a vector field in three-dimensional space, i.e. it associates a three-dimensional vector to each three-dimensional point along which the object 1 can move.

[0025] This object 1 may be any mobile object for which knowledge of the position is desired, such as a vehicle, in particular a wheeled vehicle, a drone, etc., but also a person or the body or body parts of this person (hand, head, etc.).

[0026] The object 1 is equipped with a gyrometer 11 and a magnetic measuring means 20. The means 20 and the gyrometer 11 are fixed to the object 1, for example to the bodywork in the case of a vehicle, and generally fixed in the reference frame of the object 1.

[0027] The magnetic measurement means 20 are typically magnetometers and / or magnetic gradiometers. In the latter case, instead of measuring the value of a component of the magnetic field M, the gradiometer directly measures the value of a component of the gradient of the magnetic field M, i.e. the value of the spatial derivative (optionally in addition to the magnetic field). Such magnetic gradiometers 20 are known to the person skilled in the art. According to an embodiment of the invention, it is possible to use gradiometers that directly measure the value of the second derivative (gradient of order 2), generally the i-th derivative (gradient of order i).

[0028] In the remainder of this document, the example of a magnetometer will be used, but one skilled in the art will know how to replace it with a gradiometer.

[0029] The magnetometer or magnetometers 20 are preferably "axial", i.e., the components of the magnetic field, i.e. the projection of the magnetic field vector M along their axis (or, in case of a gradiometer, the components of the spatial derivative of order i of the magnetic field, i.e. ∇ i-1 The fluctuation of the M component can be measured.

[0030] The magnetometers 20 are at least two in number, preferably at least three in number, and are capable of acquiring at least two or three components of the magnetic field, respectively. It is understood that in fact two magnetometers may be sufficient, provided that they are arranged perpendicular to each other in a plane perpendicular to the vertical axis of rotation of the vehicle, i.e. in a horizontal plane. However, the use of three magnetometers 20 and the acquisition of three components makes it possible to overcome the need for this particular provision.

[0031] Advantageously, the magnetometers 20 are advantageously organized in three groups of "triaxial", i.e. orthogonal 2x2 triplets of magnetometers 20 associated with the same spatial location and measuring the magnetic field along three axes (in the case of six magnetometers 20 there are two triaxial, in the case of nine magnetometers 20 there are three triaxial). With two triaxial it is possible to measure part of the magnetic field and its gradient (see below), and with eight appropriately positioned and oriented magnetometers 20 all the components of the gradient can be determined. This is for example the case with three triaxial, arranged at the corners of an isosceles triangle of a rectangle.

[0032] Preferably, the orthogonal system associated with the object is selected by convention (and for the remainder of this specification by the device) such that the three axes are advantageously oriented according to said orthogonal system in such a way as to facilitate calculations.

[0033] However, a person skilled in the art would know how to substitute any spatial arrangement of magnetometers / gradientometers in any case.

[0034] By "gyrometer" 11 is meant an inertial measurement means capable of measuring the angular velocity of object 1 according to a system of three orthogonal axes defining a marker of object 1, i.e. capable of measuring the three components of an angular velocity vector described as ω. It is therefore understood that gyrometer 11 can in fact designate one of three axes, in particular a set of three gyrometers associated with a three-axis arrangement (i.e. each gyrometer is capable of measuring one of the three components of the abovementioned angular velocity vector).

[0035] In the preferred embodiment where the object 1 is a vehicle, this marker preferably includes a vertical axis and two horizontal axes, in particular a longitudinal axis and a transverse axis. The rotation of the vehicle about the vertical axis is represented by the angle that the driver acts by turning the steering wheel. On a generally flat ground, the change of direction of the vehicle is in the horizontal plane, i.e. also according to said vertical axis of yaw. In practice, non-zero values ​​for roll (rotation along the longitudinal axis of the vehicle) and pitch (rotation along the transverse axis of the vehicle) are typically low, although they may be the result of, for example, an inclined road.

[0036] The method of the invention makes it possible to calibrate the three components of the angular velocity vector ω.

[0037] Furthermore, the object 1 advantageously comprises "additional" acquisition means 10 of the measured linear velocity (described as V), i.e. displacement, of the object 1. These means 10 may make it possible to obtain, directly or indirectly, the linear velocity and may therefore be of many types, for example inertial measurement means. The gyrometer 11 may thus be complemented by one or more accelerometers, so that the object 1 may also have a "six-axis" unit of inertia. The accelerometer is sensitive to external forces other than gravity acting on this sensor and is capable of measuring a certain acceleration, described as γ.

[0038] Alternatively, if the object 1 is a vehicle having wheels, the means 10 may consist of at least two odometers on the wheels, for example at least two odometers on the two rear wheels, as shown for example in FIG. 1 .

[0039] "Odometer" means a device (tachometer) that can measure the speed of a wheel by counting the number of revolutions. Generally, an odometer has a part (e.g. a magnet) attached to the wheel and detects each passing of this fixed part (the so-called "top") so as to count the number of revolutions per unit time, which is the number of revolutions. Other techniques are known, for example the optical detection of a mask on the wheel, or the magnetometer of patent FR2939514 that detects the rotation of a metal object such as a wheel.

[0040] Here, the "velocity" of the wheel is a scalar, i.e. the norm of the wheel's velocity in the ground reference system (assuming no slippage). If the radius r of the wheel is known, this norm of the velocity: v = 2πrf can be estimated by measuring the number of revolutions.

[0041] The additional means 10 may also be a GNSS receiver. It will be appreciated by those skilled in the art that the present invention is not limited to a particular technology for speed measurement.

[0042] Furthermore, the object 1 comprises the described processing means 21 (usually a processor) for the implementation, in particular in real time, of the processing of the method, for example an on-board computer of the vehicle, optionally a memory 22, and an interface 23 for returning information about the movement of the object 1 (instantaneous speed values, direction, position on a map, etc.) and / or for sending commands to the object 1. The object 1 can itself in particular be an autonomous vehicle, the processing means 21 being configured to implement the autonomous navigation of the vehicle. Said commands are thus sent to the control members of the vehicle (engine, steering wheel actuators, etc.) in order to simulate driving by a driver.

[0043] It should be noted that the processing means 21 may be external to the object 1 and may for example be connected to the object 1 by a wireless network. Alternatively, the gyrometer 11 and the magnetic measurement means 20 may be connected to the data processing means 21 by wires, for example via Ethernet.

[0044] method

[0045] The object of the method of the invention is at least a method of calibrating the gyrometer 11. By "calibration" it is meant the determination of one or more calibration parameters, a list of which is given below. In particular, certain calibration parameters can be considered reliable and predetermined, other calibration parameters are to be determined. For those to be determined, they have "current" values ​​(in other terms, the calibration has already been performed) and these values ​​are modified, if applicable (in case of a new calibration).

[0046] In a particularly preferred embodiment, the method can furthermore be a method for calibrating the magnetic measurement means 20, i.e. the gyrometer 11 and the magnetometer 20 can be calibrated simultaneously. This is a very advantageous mode, since, as will be seen, it is no longer necessary to assume that one of the gyrometry (gyrometry data) and the magnetic data (magnetic data) is reliable in order to use it as a reference for calibrating the other, i.e. the two automatically calibrate each other. Alternatively, it is of course possible to consider the magnetometer 20 as sufficiently calibrated and to calibrate the gyrometer 11 as a result, which allows, for example, to calibrate more parameters of the gyrometer 11.

[0047] Thus, in an advantageous embodiment, the method is even a method for estimating the motion of the object 1, i.e. it comprises using measurements, after calibration, to reliably estimate one or more components of the motion.

[0048] In a first step (a), the method comprises measuring the ω (measurement) gyro and obtaining, by the magnetic measurement means 20, at least two components of the magnetic field and / or the i-th derivative of the magnetic field. These components are, more precisely, the so-called measurement components, which in the case of the magnetic field are M (measurement) (For the i-th derivative, it is written as ∇ i M (measurement)Preferably, three components of angular velocity and at least six or eight magnetic components (advantageously two or three magnetic three-axis positions at the corners of an isosceles triangle of a rectangle) are obtained from which at least a part of the magnetic field and its gradients (i.e. derivatives of the following orders according to the three axes) can be derived.

[0049] These magnitudes are advantageously measured with a very small dt, typically 40 ms, prior to the characteristic time of the object 1 movement, with sampling dt (ie every "dt" seconds).

[0050] In step (b), the data processing means 21 determines a value of at least one calibration parameter of the gyrometer 11 that minimizes a first equation defined by the estimated angular velocity of the object 1 and at least one first magnetic equation of a component of the magnetic field and / or the i-th derivative of the magnetic field.

[0051] The idea is to estimate separately the gyrometric and magnetic data, which are theoretically linked by a magnetic equation, and thus thanks to this equation we can express a quantity that should ideally be zero (i.e. the gyrometric and magnetic data exactly satisfy the magnetic equation), otherwise it is the calibration that must be perfected.

[0052] ω (estimation) gyro The estimated angular velocity of the object 1, described as

number

number

[0053] In the case of three-dimensional posture, b gyro is the bias vector and D is a 3x3 matrix = (orthogonal matrix for transfer at the correct marker) x (upper triangular matrix containing scale factors and calibration). In a simplified way, for example, D is pre-defined (which in practice varies very slowly) and the only calibration parameter to be determined for the gyrometer 11 is b gyro which can be taken into account as having a tendency to vary substantially over time (called drift of the gyrometer 11).

[0054] As explained, the components of the magnetic field and / or the i-th derivative of the magnetic field obtained in step (a) are preferably measured components (i.e. potentially biased), and the components used by the first magnetic equation are estimated components, depending on the measured components and calibration parameters of the magnetic measurement means 20 (for the magnetic field the corresponding vector is M (estimation) and for the i-th derivative, it is written as ∇ i M (measurement) ).

[0055] If the magnetic measurement means 20 is a magnetometer (i.e. if the work is done on the components of the magnetic field, not on the derivatives), the estimated component M of the magnetic field (estimation) Advantageously,

number

number

[0056] b magneto matrix A generally represents "hard iron type" effects, while matrix A represents "soft iron type" effects. These effects correspond strictly to hard / soft iron effects, but can include phenomena that have the same impact but different causes (e.g., effects due to the electronics / physics of magnetometer 20).

[0057] For gyrometers more generally, the common error model

number

[0058] The fine-tuned case (

number

number

[0059] It should be noted that the magnetic measurement means 20 themselves are assumed to have been fully calibrated beforehand, and it is only their direct environment (car body, etc.) that must be corrected in the calibration to be able to estimate the geomagnetic field that would be measured in the absence of the object 1. The matrix R is therefore important and is determined so that the object magnetic markers remain stable.

[0060] Conventional calibration methods for spheres are determined only by one rotation around the nearest axis (E. Dorveaux, D. Vissiere, A. P. Martin, N. Petit, "Iterative calibration method for inertial and "magnetic sensors", in Proc. of the 48th IEEE Conf. on Decision and Control 2009). The sensor axis does not remain fixed after application of these methods.

[0061] In practice, the eight coefficients of A during calibration can be determined (one coefficient, the global scale factor that links the measurements to the physical unit of Tesla (or Gauss), cannot be determined, but at the same time, it is not necessary to know it in order to calibrate the gyrometer 11).

[0062] Thus, in a particularly preferred embodiment of the simultaneous dual calibration, there are only three calibration parameters to be determined: gyro , A and b magneto This limited number of parameters allows such a double calibration in case the object 1 has a "changing" trajectory (direction / velocity with a large diversity of values ​​as a function of time).

[0063] Magnetic Equation

[0064] The time evolution of the magnetic field at the marker of object 1 (called the body) is given by the formula

number

[0065] The first term -ω×M describes the change in the magnetic field at the marker of object 1 due to the rotation of this marker relative to the fixed marker. The second term ∇M·V describes the change in the measured magnetic field resulting from the motion of a region with a non-uniform magnetic field. Finally, the third term (outside 2) TIFF0007605460000020.tif9153 takes into account non-stationarity of magnetic fields (e.g. periodic currents like 50Hz in Europe, movement of magnets / steel objects in the vicinity of said objects, etc.).

[0066] If there are no "movable" magnetic disturbances, i.e. the magnetic field is stationary,

number

number

[0067] In the absence of any magnetic disturbances, i.e., the magnetic field is stationary and of uniform nature,

number

number

[0068] When using magnetic sensors to determine orientation, this double assumption of uniformity and stationarity is most often verified or assumed (as when using a compass). In fact, for the geomagnetic field, uniformity is locally observed as a very good approximation at our latitudes, except when metal structures or objects made of reinforced concrete (typically bridges) are present in the vicinity. Nonstationarities due to nearby displacing objects are even rarer, since the source of the disturbance must be moving, and nonstationarities due to periodic currents have zero mean and can also be handled using estimation techniques, see for example C.-I. Chesneau, M. Hillion, and C. Prieur, Motion estimation of a Rigid Body with an EKF using Magneto-Inertial Measurements, 7 thConf. on Indoor Positioning and Indoor Navigation (IPIN'16), Madrid, Spain, 2016.

[0069] It should be noted that the possibility of generalizing the magnetic equations is known in such a way that it specifically links the spatial and temporal variations of the magnetic gradient evaluated using gradiometer measurements (i.e. the spatial variations of the magnetic field) rather than the spatial and temporal variations of the magnetic field evaluated using magnetometer measurements. More precisely, the equation

number

number

[0070] In the remainder of this specification, depending on the product, zero-order equations are used (i.e.,

number

[0071] With reference to FIG. 2, the present calibration method elegantly proposes a calibration with several levels of assumptions.

[0072] More precisely, in step (b), the gyrometer is first calibrated by making the dual assumptions of homogeneity and stationarity of the magnetic field, and then, in an optional step (c) of estimating parameters representative of the errors on the calibration parameters, the absence of disturbances, i.e. the validity of the assumptions, is verified.

[0073] In a preferred embodiment, if this verification fails (said parameter representing the error is greater than a predefined threshold), the calibration is attempted again in step (d) by making only the assumption of stationarity (and no longer the assumption of uniformity). In step (e), the validity of this assumption can be re-verified. Alternatively, if the verification of step (c) fails (said parameter representing the error is greater than a predefined threshold), the method can be stopped and the current calibration maintained.

[0074] However, it is understood that the presence of step (d) makes it possible to carry out a calibration in cases where the magnetic field may not be homogeneous, provided that it has the aforementioned additional collecting means 10. However, the invention is not limited to this embodiment and it is possible to have only the gyrometer 11 and the magnetic measuring means 20 and to carry out only steps (a), (b) and optionally (c).

[0075] A possible step (d) therefore comprises the acquisition by additional acquisition means 10 of the measured linear velocity of the object 1, described as V (in practice this acquisition can be carried out simultaneously with the other acquisitions of step (a) in particular at each time step), and then redetermining by the data processing means 21 the value of at least one calibration parameter of the gyrometer 11 which minimizes a second equation defined by the estimated angular velocity and linear velocity of the object 1 and at least one second magnetic equation for the components of the magnetic field and / or the i-th derivative of the magnetic field, the at least one second magnetic equation only assuming that the magnetic field is stationary around the magnetic measurement means 20.

[0076] "Obtaining" is to be understood in a broad sense, i.e. directly or indirectly. In practice, the means 10 measure first magnitudes (acceleration, wheel revolutions, etc.) which are most often not linear velocities, and the so-called measured linear velocity is in fact deduced from these magnitudes and optionally combined. The skilled person will know how to obtain the linear velocity of the object 1 in many ways, for example by magnetic gradiometry.

[0077] Therefore, at least one second equation, which only assumes that the magnetic field is stationary around the magnetic measuring means 20 used in step (d), preferably reads:

number

number

[0078] Therefore, under the assumption of stationarity of the magnetic field around the magnetic measuring means 20, the quantity

number

number

[0079] Similarly, the at least one first magnetic equation, which assumes that the magnetic field used in step (b) is uniform and stationary around the magnetic measuring means 20, is a simplified version of the at least one second equation, preferably:

number

number

number

number

[0080] The first calibration proves to be easy to perform, since it does not require knowledge of the linear velocity V.

[0081] Advantageously, the first equation (which is minimized in step (b)) is

number

number

number

number

number

number

[0082] To implement this minimization, the data processing means 21 can operate over time on intervals of a given length, as such, in known manner, recursive filters (RLS, recursive least squares, etc.) or optimization (least squares, etc.) can be used.

[0083] For example, the calibration parameters of the magnetometer 20, i.e., parameters A and b magneto On the assumption that are simultaneously determined,

number

number

number

number

[0084] Error Characterization

[0085] As explained, the method advantageously includes a step (c) of estimating parameters representative of the errors on the calibration parameters and, advantageously, if steps (c) and (d) have been performed, a step (e) of newly estimating said parameters representative of the errors on the calibration parameters.

[0086] The idea is to estimate the quality of the information provided by the magnetic measurement means 20 in order to verify or not verify the level of assumptions about the ambient magnetic field. More precisely, as shown in FIG. - if the first estimate of the parameter representative of the error (step (c)) is smaller than said predefined threshold, the double assumption of homogeneity and stationarity of the ambient magnetic field is verified and the result of the calibration of step (b) is accepted. Step (c) then comprises an effective calibration of the gyrometer 11 and / or the magnetic measuring means 20 by the values ​​determined in step (b) of the calibration parameters, and there is no need to carry out steps (d) and (e); if the first estimate of the parameter representative of the error (step (c)) is greater than said predefined threshold, the double assumption of homogeneity and stationarity of the ambient magnetic field is rejected and the result of the calibration of step (b) is not accepted and one stops, or step (d) in which: o if the second estimation of the parameter representative of the error (step (e)) lasts for a time less than said predefined threshold, the assumption of the sole stationarity of the ambient magnetic field is verified and the result of the calibration of step (b) is accepted. Step (e) then comprises a valid calibration of the gyrometer 11 and / or the magnetic measuring means 20 by the value determined in step (d) of the calibration parameter; If the second estimate of the parameters representative of the error (step (e)) is always greater than said predefined threshold, even the assumption of stationarity of the ambient magnetic field is rejected (perturbations are present) and the results of the calibration of step (d) are not accepted. The parameters of the former calibration are therefore kept as the calibration. Then the complete magnetic equation (terms (outside 2) It is possible to try either to use a different magnetic field (containing TIFF0007605460000046.tif9153) or to temporarily discard the magnetic measurements; Step (d) of recalibration can be performed:

[0087] It should be noted that the determined values ​​during the occurrence of steps (b) and / or (d) are not used for the valid calibration, but can be stored in the data storage means 12 and can be used during future occurrences of steps (c) and / or (e). For example, as long as the parameter representing the error exceeds a threshold, the determined calibration parameter is stored, and if it is below the threshold, the valid calibration also takes into account the stored value. Furthermore, the outcome of steps (c) and (e) may depend on the age of the previous calibration parameter (for example via the value of the threshold).

[0088] It should be noted that the accumulated error interval after a duration t since the last good calibration can be estimated according to this duration t, e.g., the estimated error in this calibration by adding to this error an increment proportional to the duration t. The predicted error interval can be used to characterize the result of future calibrations, and in particular to define the value of said predetermined threshold: e.g., the larger the error interval, the more urgent it is to recalibrate the gyrometer, and therefore the more "tolerant" is the value of the parameter representing the error (i.e., the higher the predetermined threshold). Furthermore, the error interval increases proportionally faster when the estimated error is large, which makes it possible to hold "good" calculations with very low estimated errors for a long time, while forcing "not-so-good" calibrations to be recalibrated faster.

[0089] Generally, the parameter representing the error on the calibration parameter is a function of the value of the first / second equation calculated for the determined value of the calibration parameter (depending on whether it is the first estimated occurrence or the second estimated occurrence of the parameter).

[0090] The first embodiment of steps (c) / (e) is said to be internal since it uses only the magnitudes available in steps (b) / (d). Preferably, the estimated residuals from steps (b) / (d) are used, i.e. said parameters representing the errors are in particular the norms of the values ​​of the first / second equations over a given time interval (e.g. L 2 Or L ∞), typically

number

number

[0091] Alternatively, the deviation can be quantified relative to a normal distribution, i.e.

number

number

[0092] In a second embodiment of steps (c) / (e), called extrinsic, another type of magnitude is used in such a way that a so-called reference angular velocity is calculated based on one of these magnitudes to compare with the estimated angular velocity. The person skilled in the art knows how to calculate such a reference angular velocity from various kinds of magnitudes, in particular, for example: - Steering wheel angle (if object 1 is a vehicle), - GNSS data, e.g. GPS (if / when available), - Odometer data, It is possible to use

[0093] In the latter case, if the additional means 10 for acquiring the linear velocity consist of at least two odometers (if the object 1 is a vehicle), said theoretical angular velocity can in fact be obtained from the measured velocities of the wheels. In particular, if the rear wheels are each equipped with an odometer, assuming an Ackermann differential geometry, the second estimated angular velocity of the object 1 is given by:

number

number

[0094] According to a third embodiment, for the first generation of the estimation of the error parameters (step (c)), all or part of the spatial gradient of said component of the magnetic field and / or the i-th derivative of the magnetic field, i.e. ∇M or ∇ i+1 M can be estimated quite simply according to the order of the second equation used.

[0095] In practice, this estimate must be close to zero to maintain the uniform nature. It will be appreciated that an estimate of a subset of the gradient's components is sufficient for this characterization in practice, since each non-zero component is already a manifestation of non-uniformity.

[0096] Alternatively or supplementally, learning can be used to improve the estimation of this error parameter and / or to develop approaches for identifying favorable conditions (i.e., stationary and uniform) in a more robust manner and with increased availability.

[0097] In particular, it is possible to implement learning mechanisms such as neural networks, machines using support vectors, nearest neighbor methods, decision trees, etc. Thus, at each occurrence of steps (b) to (e), a learning base can be enriched in which each measurement data set is "tagged" with the corresponding value of the parameter representative of the error, in such a way that the calibration learns progressively (as successive occurrences of steps (b) to (e) occur) and automatically to distinguish acceptable calibrations from unacceptable calibrations. In this way, the calibration is constantly improved.

[0098] Movement Estimation

[0099] As explained, the method advantageously comprises a step (f) of estimating by the data processing means 21 the motion of the object 1 depending on the measured angular velocity of the object 1 and / or on the magnetic field and / or on at least two components of the i-th derivative of the magnetic field and / or on any measured linear velocity of the object 1 (by the means 10) and on the values ​​of the calibration parameters, i.e. after recalibration. This step (f) can be performed in parallel and successively, although it is not shown in Fig. 2 which focuses on the calibration.

[0100] By "motion estimation" is meant in particular an estimation of at least the orientation (horizontal plane, i.e., heading) of the object 1 and advantageously an estimation of the normal velocity. The orientation is typically obtained by integration of the angular velocity from a known initial orientation.

[0101] Preferably, the estimation by the data processing means 21 of the orientation of said object 1 is performed only depending on the measured angular velocity of the object 1 and the values ​​of the calibration parameters, i.e. using only the gyrometric data. In summary, the magnetic data is used to calibrate the gyrometer 11 and then only the gyrometer data is used to estimate the orientation. If applicable, the measured linear velocity (by the optional means 10) is only used to determine the global velocity of the object 1 in step (f).

[0102] It should be noted that in the case of an autonomous vehicle, step (f) may include generating commands for said vehicle 1 in response to the estimated motion, for example in such a way as to bring the vehicle 1 to a desired destination or to stop the vehicle 1 by keeping it on an obstacle-free trajectory.

[0103] Equipment and objects

[0104] According to a second aspect, the invention relates in particular to a set of devices 11, 20, 21 and optionally 10 for carrying out one or another of the embodiments of the method.

[0105] This set can be installed as a kit in a "conventional" object 1 in such a way as to transform it. Alternatively, in particular if the object 1 is a vehicle, this can be an autonomous vehicle already equipped with data processing means 21 for navigation of the vehicle as well as sensors such as a gyrometer 11 and / or an odometer used as additional acquisition means 10.

[0106] The object 1 is proposed in particular for wheeled vehicles and includes: - a gyrometer 11 configured to obtain a measured angular velocity of the object 1; - magnetic measurement means 20 configured to obtain at least two components of the magnetic field and / or the i-th derivative of the magnetic field; - if applicable, additional acquisition means 10 configured to acquire a measured linear velocity of the object 1 (advantageously odometers mounted on at least two of the wheels of the vehicle and configured to acquire the measured velocities of said two wheels); - a data processing means 21, configured to determine values ​​of at least one calibration parameter of the gyrometer 11 that minimize a first equation defined by an estimated angular velocity of the object 1 and at least one first magnetic equation of a component of the magnetic field and / or an i-th order derivative of the magnetic field, the estimated angular velocity of the object 1 is a function of the measured angular velocity and the calibration parameters of the gyrometer 11; at least one first magnetic equation assumes that the magnetic field is uniform and stationary around the magnetic measuring means 20; The data processing means 21 may further be configured as follows: o Estimate parameters that represent the errors on the calibration parameters: o If the parameter representing the error is greater than a predefined threshold, redetermine the value of at least one calibration parameter of the gyrometer 11 that minimizes the estimated angular velocity of the object 1 and a second equation defined by at least one second magnetic equation of the components of the measured linear velocity, the magnetic field and / or the i-th derivative of the magnetic field, the at least one second magnetic equation only assuming that the magnetic field is stationary around the magnetic measurement means 20.

[0107] As mentioned above, the object 1 may further include a memory 22 and an interface 23, as well as other sensors such as a steering angle sensor or a GNSS receiver.

[0108] Furthermore, the data processing means 21 can be configured to estimate the movement of the object 1 according to the measured angular velocity of the object 1 and / or the measured linear velocity of the object 1 (if applicable, the result of comparing said parameter representing the error with a threshold value) and the value of the calibration parameter.

[0109] Computer Program Products

[0110] According to a third and fourth aspect, the present invention relates to a computer program product comprising code instructions for the execution (on a processing means 21) of the method for calibrating a gyrometer 11 according to the first aspect of the invention, and to a storage means readable by a part of a computing device (e.g. a data storage means 22) in which this computer program product is found.

Claims

1. 1. A method of calibrating a gyrometer of an object moving in an ambient magnetic field, the method comprising: (a) an obtaining step, of the measured angular velocity of the object by the gyrometer, and at least two measured magnetic components which are components of the magnetic field and / or the i-th spatial derivative of the magnetic field around the magnetic measurement means, i being an order of the spatial derivative of the magnetic field, i≧1, by a magnetic measurement means fixed to the object; The steps of obtaining; (b) determining by the data processing means at least one candidate value for at least one calibration parameter of the gyrometer that minimizes the value of a first equation defined by at least one first magnetic equation relating the estimated angular velocity of the object and a component of the magnetic field and / or a representative magnetic component representing an n-th order spatial derivative of the magnetic field around the magnetic measurement means, where n indicates an order of spatial derivative of the magnetic field, n+1≧i and n≧1; the estimated angular velocity of the object is a function of the measured angular velocity and a calibration parameter of the gyrometer; said representative magnetic component being a function of said measured magnetic component, said at least one first magnetic equation assumes that the magnetic field is uniform and stationary around said magnetic measurement means; The step of determining (c) estimating a first error parameter representative of an error on the at least one candidate value; A method comprising:

2. The estimated angular velocity ω of the object （ｅｓｔｉｍａｔｉｏｎ） ｇｙｒｏ is the model [0010] The measured angular velocity ω （ｍｅａｓｕｒｅｍｅｎｔ） ｇｙｒｏ Linked to D and b ｇｙｒｏ is the calibration parameter of the gyrometer, The method of claim 1.

3. if the first error parameter is greater than a predefined threshold, determining by the data processing means at least one updated candidate value for the at least one calibration parameter of the gyrometer that minimizes a value of a second equation defined by at least one second magnetic equation relating the estimated angular velocity of the object and the representative magnetic component, the at least one second magnetic equation assuming that the magnetic field is stationary around the magnetic measurement means; The method according to claim 1 or 2.

4. The step (d) includes acquiring a measured linear velocity of the object by an additional acquisition means, and the at least one second magnetic equation relates the estimated angular velocity of the object, the representative magnetic component and the measured linear velocity, The method according to claim 3.

5. The second formula is: [0025] is a function of ∇ ｎ M is the vector of the representative magnetic component, ω is the estimated angular velocity, and ∇ ｎ＋１ M is the first spatial derivative of the vector of the representative magnetic component, V is the measured linear velocity, and f ｎ and g ｎ is a given function, The method according to claim 4.

6. (e) estimating a second error parameter representative of an error on the at least one updated candidate value.

6. The method according to any one of claims 3 to 5.

7. the second error parameter is a function of the estimated angular velocity calculated using the at least one updated candidate value for the at least one calibration parameter of the gyrometer. The method according to claim 6.

8. The second error parameter is [0030] is the average of ∇ ｎ M is the vector of the representative magnetic component, ω is the estimated angular velocity, and ∇ ｎ＋１ M is the first spatial derivative of the vector of the representative magnetic component, V is the measured linear velocity, and f ｎ and g ｎ is a given function, The method according to claim 7, which relies on claim 6, which relies directly or indirectly on claim 4.

9. the at least one calibration parameter of the gyrometer has at least one current value determined in a previous calibration of the gyrometer, and step (c) includes determining a previous error parameter representative of an error on the current value and a cumulative error interval that is a function of a duration since a last calibration of the gyrometer, and the predetermined threshold is dependent on the cumulative error interval.

9. The method according to any one of claims 3 to 8.

10. the representative magnetic component is an estimated component which is a function of the measured magnetic component and a calibration parameter of the magnetic measurement means, and step (b) also comprises determining at least one additional candidate value for at least one calibration parameter of the magnetic measurement means, and step (d) also comprises determining at least one updated additional candidate value for the at least one calibration parameter of the magnetic measurement means.

10. The method according to any one of claims 3 to 9.

11. The first error parameter is: the estimated angular velocity calculated using the at least one candidate value for the at least one calibration parameter of the gyrometer; or all or part of the spatial gradient of the measured magnetic component; is a function of 11. The method according to any one of claims 1 to 10.

12. the representative magnetic component is an estimated magnetic component which is a function of the measured magnetic component and a calibration parameter of the magnetic measurement means, and step (b) also comprises determining at least one additional candidate value for at least one calibration parameter of the magnetic measurement means.

12. The method according to any one of claims 1 to 11.

13. The magnetic measurement means is a three-axis magnetometer, and the estimated magnetic component is modeled [0045] to the measured magnetic component by M （ｅｓｔｉｍａｔｉｏｎ） corresponds to the estimated magnetic component, and M （ｍｅａｓｕｒｅｍｅｎｔ） corresponds to the measured magnetic component, and A and b ｍａｇｎｅｔｏ is the calibration parameter of the magnetic measurement means, The method of claim 12.

14. The first magnetic equation is [0050] , which is of the form ∇ ｎ M is the vector of the representative magnetic component, ω is the estimated angular velocity, f ｎ is a given function, 14. The method according to any one of claims 1 to 13.

15. The first error parameter is [006] is the average of The method of claim 14.

16. and (f) estimating, by the data processing means, a motion of the object according to the measured angular velocity of the object, the measured magnetic component and the calibration parameters, wherein an orientation of the object is estimated in step (f) solely as a function of the measured angular velocity of the object and the calibration parameters.

16. The method according to any one of claims 1 to 15.

17. An object moving in an ambient magnetic field, said object comprising a gyrometer configured to obtain a measured angular velocity of said object, and magnetic measurement means configured to obtain at least two measured magnetic components which are components of said magnetic field and / or an i-th spatial derivative of said magnetic field, where i indicates an order of said spatial derivative of said magnetic field, i≧1, said object comprising a data processing means comprising: configured to determine at least one candidate value for at least one calibration parameter of the gyrometer that minimizes the value of a first equation defined by at least one first magnetic equation relating said component of the magnetic field and / or a representative magnetic component representing an n-th order derivative of said magnetic field around the magnetic measurement means, where n denotes an order of spatial differentiation of said magnetic field, n+1≧i and n≧1; the estimated angular velocity of the object is a function of the measured angular velocity and the calibration parameters of the gyrometer; the representative magnetic component is a function of the measured magnetic component, said at least one first magnetic equation assumes that the magnetic field is uniform and stationary around said magnetic measurement means; - configured to estimate a first error parameter representative of an error on said at least one candidate value; Further comprising a data processing means. object.

18. the object further comprises additional acquisition means configured to acquire a measured linear velocity of the object, and the data processing means is further configured to determine, if the first error parameter is greater than a predefined threshold, at least one updated candidate value for the at least one calibration parameter of the gyrometer that minimizes a value of a second equation defined by at least one second magnetic equation relating the estimated angular velocity of the object, the measured linear velocity and the representative magnetic component, the at least one second magnetic equation assuming that the magnetic field is stationary around the magnetic measurement means.

20. The object of claim 17.

19. the object is a vehicle having wheels, and the additional acquisition means are at least two odometers attached to the wheels of the vehicle; 20. The object of claim 18.

20. A computer program product comprising code instructions for carrying out, when executed on a computer, a method for calibrating a gyrometer according to any one of claims 1 to 16.

21. 17. A storage means readable by a part of a computing device, said storage means comprising a computer program product comprising code instructions for carrying out a method for calibrating a gyrometer according to any one of claims 1 to 16.

Images