FR3098295A1 - Improved inertial sensor - Google Patents

Abstract

A method (100) for determining a quadrature command (CTq) and a frequency command (CTf) of a vibration wave generated by a resonator (Res) of an inertial angular sensor (10), the method comprising the steps of: -A determining the electrical angle (θ), -B estimating first values (Kq ', ΔK') of the quadrature and equalization stiffnesses from a first (TrimQ) and d respectively 'a second (TrimF) slaving, said first values being estimated in the wave frame X'Y', -C determine second values (Kq, ΔK) of the quadrature and equalization stiffnesses in the sensor frame XY, from the first values of said stiffnesses (Kq ', ΔK') estimated in step B, -D determine the quadrature command (CTq) and the frequency command (CTf) corresponding respectively to said second values (Kq, ΔK) determined at l ' step C, -E apply the frequency command (CTf) and the quadrature command (CTq) determined in step D. e for the abstract: Fig. 5

Description

Improved inertial sensor

FIELD OF THE INVENTION

The field of the invention is that of vibrating inertial sensors in which two masses are vibrated. More particularly, the invention relates to inertial sensors of the MEMs type having a planar structure, typically micro-machined in a support plate.

STATE OF THE ART

Tuning fork inertial sensors are known to those skilled in the art. Inertial sensors micro-machined in a thin flat plate, allowing the measurement of an angular position (gyroscope) or an angular speed (gyrometers) are described in the document EP2960625. We recall the main characteristics below.

The manufacture of these micro-machined sensors, also known as MEMS (Micro-Electro-Mechanical-Systems) sensors, uses collective micro-machining techniques, etching, doping deposits, etc., similar to those used for the manufacture of circuits integrated electronics, allowing low production costs.

These sensors consist of two mobile vibrating masses M1 and M2 illustrated arranged one around the other (concentric) and excited in vibration in tuning fork mode in the plane of the plate (XY plane in the figure) via one or more excitation transducers. The two masses are suspended from fixed anchor points A of the plate by suspension springs RS (orthotropic). The two masses are coupled together by RC stiffness elements. The aim is to obtain by construction a stiffness along X equal to a stiffness along Y and a coupling stiffness between X and Y that is zero. The useful mode of vibration corresponds to a linear vibration of the two masses in phase opposition.

This architecture forms a resonant system with two masses coupled together by the Coriolis acceleration. When the gyrometer rotates around the Z axis perpendicular to the XY plane called the sensitive axis, the composition of the forced vibration with the angular rotation vector generates, by Coriolis effect, forces which put the mobile masses in natural vibration perpendicular to the vibration d excitation and sensitive axis; the amplitude of the natural vibration is proportional to the speed of rotation. The electronics associated with the sensor calculate the amplitude of the vibration according to the direction orthogonal to the direction of excitation whatever the latter (known by hypothesis).

The sensor can operate in gyrometer mode: the direction of natural vibration is kept fixed relative to the sensor case by modifying the excitation and the output information is then an image of the necessary energy that must be applied to the transducers. excitation to keep the direction of natural vibration fixed despite the movements of the case. The measurement of this counterforce gives access to the angular speed Ωof the sensor. The sensor can also operate in gyroscope mode: the direction of the natural vibration is left free and is detected to give the angular orientation of the sensor.

The gyrometer mode has the following advantages: (i) there is no angular noise, more precisely no angle error linked to the position of the vibration and (ii) there are no variations in angular speed errors (drift ) related to the angle (by definition because we keep a constant angle with respect to the reference of the gyrometer).

The gyroscope mode has the following advantages: (i) having a very reduced scale factor error compared to the gyrometer mode and (ii) having a very high operating dynamics.

The entire structure of the resonator is axisymmetric around two axes X and Y defining a sensor reference as shown . By axisymmetric is meant that the structure is symmetrical with respect to X and symmetrical with respect to Y. As described below, these axes constitute the main directions of the actuators/detectors, which operate along these two axes.

To excite the useful mode of vibration in any given direction of the plane, the excitation signal is broken down into two components of respective adjusted amplitudes, applied respectively to the excitation transducer Ex acting in the direction X and to the excitation transducer Ey acting in direction Y associated with at least one mobile mass (internal mass M1 on the ). These transducers are able to maintain a forced vibration via an amplitude command (to fight against the damping of the MEMs) and in any direction of the XY plane, via a precession command (to make the wave rotate).

The movements of the resulting wave are detected by combining the information collected by at least one pair of detection transducers Dx, Dy recovering the position of the mass in its course in the sensor reference XY (two of each on the ) and associated with at least one mobile mass.

The transducers are preferably made by interdigital comb electrodes with air gap variation. There is a fixed comb whose teeth are integral with a fixed electrode of the machined plate and a mobile comb whose teeth, interdigitated with the teeth of the fixed comb, are integral with the mobile mass associated with the transducer considered.

The excitation consists in applying an alternating voltage between the mobile comb and the fixed comb, at the desired vibration frequency (mechanical resonance frequency of the suspended mobile mass). The movement generated is perpendicular to the teeth of the comb.

The detection consists in applying a bias voltage between the fixed comb and the mobile comb and in observing the variations in charge which result from the variations in capacitance between the fixed comb and the mobile comb due to the variations in spacing between the teeth of the fixed comb. and moving comb. The movement measured is the movement perpendicular to the teeth of the comb.

It is well known to those skilled in the art that imperfections in the production of the sensor lead to errors in the information delivered at the output thereof. Most of these imperfections must be compensated for by balancing the gyrometer.

It is known to achieve this compensation by locally removing material, for example by laser ablation to modify the distribution of mass or stiffness. This method is expensive or even impossible to implement on a micro-machined gyrometer in a thin silicon plate, whose detection and excitation movements are located in the plane of the substrate.

Two types of imperfections are distinguished in terms of stiffness. The mass/spring vibrating assembly is characterized by a 2x2 stiffness matrix. This matrix is symmetrical, characterized in the reference XY by Kx stiffness according to X, Ky stiffness according to Y and Kxy stiffness of coupling between X and Y (with Kyx=Kxy). Due to production imperfections, Kx is different from Ky and Kxy is non-zero, whereas for optimum operation of the sensor, we seek to obtain Kx=Ky and Kxy=0, i.e. a matrix of proportional final stiffness to identity.

The axis of vibration of the wave is called the X' axis. This axis defines a reference X'Y', with Y' perpendicular to X' in the plane of the MEMs. The X' axis forms with the X axis an angle called the electrical angle θ and the X'Y' frame is called the wave frame. We will assume for the moment that the wave vibrates according to X (X'=X).

The first type of imperfection is the frequency difference between the main axis of vibration and the axis perpendicular to the vibration in the plane of the MEMs, corresponding to a stiffness matrix of the system in which the stiffness along the X axis is different from the stiffness along the Y axis. It is sought to equalize the resonance frequencies along the two aforementioned axes by means of an adjustable electrostatic stiffness. This electrostatic stiffness, called equalization stiffness, is delivered by frequency adjustment transducers Tx, Ty (at least one pair on at least one ground see ) acting along the X and Y directions. Its application aims to equalize the stiffnesses along the two axes of the vibration, by reducing the value of the highest stiffness, thus making the frequencies equal. The frequency correction is called “frequency trimming” in English.

A second type of imperfection comes from the mechanical coupling between the axis of the vibration and the perpendicular axis, at the origin of the so-called quadrature bias. It is about the defects of anisotropy of dynamic stiffness of the whole of the two vibrating masses, resulting in a vibration either linear but elliptical, and corresponding to the existence of a stiffness of coupling Kxy not null. A solution is to cancel this term by applying a force F (sinusoidal) to the system via the excitation transducers. The problem is that the application of this force is not exerted at exactly the right time (phase errors) and in the right axis (gain error), causing the application of drifts. To avoid applying a force F, we physically cancel the term Kxy not by applying a force but by directly changing the stiffness of the resonator via at least one pair of transducers Q+ and Q- as shown (2 Q+/Q- pairs on the ). These transducers operating on X and Y are arranged on the diagonals to respect the symmetry and the “geometrical” anisotropy, and for space reasons. Correcting for quadrature bias is referred to as quadrature trimming.

Thus the transducers of the quadrature "trim" modify the characteristics of the MEMs sensor to eliminate the couplings between the two axes of the wave frame and the transducers of the frequency "trim" modify the characteristics of the MEMs sensor to eliminate the frequency differences between the two axes of the wave marker.

The transducers Tx, Ty, Q+ and Q− are preferably also interdigital combs as illustrated in FIGS. 2 and 3, controlled by DC voltages.

Preferably, excitation, detection, frequency adjustment and quadrature bias correction transducers are produced on the two masses, as shown , the index 1 corresponding to the mass M1 and the index 2 to the mass M2. Figures 2 and 3 are non-limiting examples of layouts, many other types of layouts are possible, with the constraint of producing an axisymmetric system.

There illustrates the operation of an inertial sensor CI0 according to the state of the art, and more particularly the servo-controls of the frequency and quadrature “trim”. The Res resonator includes the various transducers described above and symbolized by E (excitation), D (detection), TQ (quadrature trim) and TF (frequency trim). The wave of vibration OV vibrates according to X (X'=X). The vibration modes along X'Y' coincide with the excitation and detection modes.

Three drives drive the excitation combs E in parallel: a precession drive (not shown) maintains the vibration wave at a predetermined angle (measurement of the reaction force which counters the Coriolis force in gyrometer mode); an amplitude servo (not shown) keeps the wave vibration constant and a quadrature servo controls the force fy to keep the vibration linear (via the Ctqe command). An issue with the quadrature fy force is the phase and gain or direction accuracy of its application. A fourth PLL type loop (not shown) seeks to identify the phase of the oscillation (position of the mass in its travel). This PLL loop does not influence the vibration wave, it acts as an observer. Thanks to the information delivered by the phase loop, it is possible to position with the correct phase the forces to be sent to Ex and Ey via the three aforementioned servocontrols and carry out the demodulation of the detection signals.

The detection transducers measure the position of the vibration (x,y) in the XY sensor frame. In addition to the three enslavements on the excitation, a first and a second enslavement are implemented respectively for the quadrature trim and the frequency trim. A processing unit UT performs the various calculations and generates, for the corrections, commands to the various transducers: a command CTqe to apply a quadrature force via the transducer E, a frequency trim command CTf for TF, a trim command quadrature CTq for TQ. The CTq and CTf commands of the trimmers are DC voltages which modify the intrinsic characteristics of the resonator, while the CTqe command of the force is a sinusoidal voltage (see above). The quadrature control CTq on TQ is set to obtain, in steady state, zero quadrature force (CTqe control) applied to the excitation E, which solves the problems related to the application of this force.

The signal coming from y is demodulated in cosine and in sine. Cosine demodulation is used for the first servo relative to quadrature trim. The resulting signal is processed by a corrector Coq1 and delivers an estimate of the quadrature stiffness Kq intended to cancel the coupling stiffness Kxy. After a second corrector Coq2 (integrator) and the transformation of the stiffness into electrical voltage by the device Gq, the quadrature command CTq is applied to TQ. Sine demodulation is used for the second frequency trim servo. The resulting signal is processed by a corrector Cof which generates the equalization stiffness ΔK, then a device Gf transforms the stiffness into an electrical voltage so as to generate the frequency trim command CTf.

The drives for frequency and quadrature correction were initially developed for non-axisymmetric sensors (X and Y do not play the same role) configured to operate with a wave vibrating along X. In this case the first and second drives operate independently and function properly.

For an axisymmetric sensor which allows the use of a vibrating wave at an angle θ different from 0°, the trims become interdependent and no longer work correctly. For example for certain angles such as θ = π/4 the quadrature trim no longer has any effect on the coupling stiffness Kxy and generates instability of the first and second servo-controls. The instability described above leads to saturation of the actuators and the sensor must be restarted. Thus, when operating the MEMs sensor at an angle θ other than zero, it may be preferable to cut the trim servos, which leads to sensor measurement errors related to the application of additional forces.

An object of the present invention is to remedy the aforementioned drawbacks by proposing a mode of operation of the sensor allowing effective implementation of the frequency correction servocontrols and of the quadrature bias for a wave vibrating at any electrical angle.

DESCRIPTION OF THE INVENTION

The subject of the present invention is a method for determining a quadrature command and a frequency command of a vibration wave generated by a resonator of an inertial angular sensor, the resonator having a planar and axisymmetric structure around two mutually perpendicular X and Y axes defining an XY sensor frame and comprising two vibrating mobile masses (M1, M2) arranged one around the other, coupled together by coupling springs and configured to vibrate in phase opposition along a direction X' defining a wave mark X'Y', the resonator further comprising a plurality of electrostatic transducers controlled by electric voltages and operating along the two axes X and Y, at least of which on at least one of the two masses the following transducers:

A pair of excitation transducers referred to as E transducers configured to hold the wave at a constant amplitude via an amplitude control (Ca) and, if necessary, to rotate said vibrational wave via a precession control (Cp) , a pair of sensing transducers referred to as D transducers configured to detect movement of the vibration wave, a pair of quadrature bias compensating transducers referred to as TQ transducers, configured to apply quadrature stiffness via quadrature control (CTq), the quadrature stiffness being configured to cancel a coupling stiffness between X' and Y', and a pair of frequency adjusting transducers referred to as TF transducers, configured to apply equalizing stiffness via frequency control (CTf), the equalization stiffness being configured to cancel a stiffness difference between X' and Y' so as to equalize the resonance frequencies of the vibration wave. on on X' and Y'.

The method applies when the inertial sensor is in operation with a vibration wave vibrating along X' characterized by an electrical angle, the method comprising the steps consisting in:
-To determine the electrical angle,
-B estimate first values of said quadrature and equalization stiffnesses from respectively a first and a second servo-control, said first values being estimated in the wave frame X'Y',
-C determining second values of said quadrature and equalization stiffnesses in the XY sensor frame, from the first values of said stiffnesses estimated in step B,
-D determine the quadrature command and the frequency command corresponding respectively to said second values determined in step C,
-E apply the frequency command (CTf) and quadrature command (CTq) determined in step D.

According to a variant, the inertial sensor operates in gyrometer mode, the electrical angle determined in step A being equal to an angle imposed via the precession command.

According to another variant, the inertial sensor operates in gyroscope mode, the electrical angle resulting from a rotation of the inertial sensor being measured by said inertial sensor, the electrical angle determined in step A being equal to said measured angle of rotation.

According to yet another variant, the method according to the invention comprises:
- a first phase in which the electrical angle describes a plurality of electrical angles obtained by applying said precession command, steps A to E being implemented for each electrical angle, step D further comprising a sub-step of storage of the associated frequency command value and a sub-step of determining a law of variation of the frequency command as a function of the electrical angle,
- a second phase in which the inertial sensor operates in gyroscope mode, the electrical angle left free resulting from a rotation of the inertial sensor and being measured by said inertial sensor, the second phase comprising:
*a step (BO) of open-looping the second servo-control, the frequency command applied then being determined from said law of variation for said measured angle of rotation,
*a step of detecting a resonance frequency deviation, the open looping step being implemented as long as said resonance frequency deviation is less than or equal to a predetermined threshold,
*a closed-loop reset step of said second slaving when the frequency difference is greater than said threshold, the method then looping back to the first phase for an update of said variation law.

According to one embodiment, step B comprises a sub-step B1 of determining a position of the vibration wave in the wave reference X'Y' from the measurement of a position of the vibration wave in the sensor reference XY and of the electrical angle, and a sub-step B2 of estimating the first values of said quadrature and equalization stiffnesses from said position in the wave reference.

According to one embodiment, step C consists in determining a vector defined by said second values by applying a rotation by an angle equal to twice the electrical angle to the vector defined by said first values.

According to another aspect, the invention relates to an inertial angular sensor comprising:
-a resonator having a planar and axisymmetric structure around two mutually perpendicular X and Y axes defining an XY sensor marker, and comprising two vibrating mobile masses arranged one around the other, coupled together by coupling springs and configured to vibrate in phase opposition according to a wave of vibration vibrating in a direction X' characterized by an electrical angle and defining a wave marker X'Y', the resonator further comprising a plurality of electrostatic transducers controlled by electrical voltages and operating along the two axes X and Y, including at least, on at least one of the two masses, the following transducers:

A pair of excitation transducers referred to as E transducers configured to hold the wave at a constant amplitude via amplitude control and, if required, to rotate said vibrational wave via precession control, a pair of sensing referred to as D transducers configured to sense movement of the vibration wave, a pair of quadrature bias compensating transducers referred to as TQ transducers configured to apply a quadrature stiffness via a quadrature drive, the quadrature stiffness being configured to cancel a coupling stiffness between X' and Y', and a pair of frequency adjustment transducers referred to as TF transducers, configured to apply an equalizing stiffness via a frequency command, the equalizing stiffness being configured to cancel a difference in stiffness between X' and Y' so as to equalize the resonance frequencies of the vibration wave on X' and Y'. The quadrature and equalization stiffnesses are determined from a first and a second servo-control, respectively.

The sensor further comprises a processing unit configured to determine said electrical angle and comprising:
-a first module configured to estimate first values of said quadrature and equalization stiffnesses from the first and second servo-control respectively, said first values being estimated in the wave frame X'Y',
-a second module configured to determine second values of said quadrature and equalization stiffnesses in the XY sensor frame, from the first values of said stiffnesses,
-a set of two electrical gain modules configured to respectively determine the quadrature command corresponding to the second value of the quadrature stiffness and the frequency command corresponding to said second value of equalization stiffness,
the transducers TF and TQ being configured to respectively apply said frequency command and said quadratureau resonator command.

According to one embodiment, the first module is configured to determine a position of the vibration wave in the wave reference X'Y' from the electrical angle and the measurement of a position of the vibration wave in the sensor reference XY produced by the transducers D, and to estimate first values of said quadrature and equalization stiffnesses from said position in the wave reference.

According to one embodiment, the second module is configured to determine a vector defined by said second values by applying a rotation by an angle equal to twice the electrical angle to the vector defined by said first values.

The following description presents several embodiments of the device of the invention: these examples do not limit the scope of the invention. These exemplary embodiments present both the essential characteristics of the invention as well as additional characteristics related to the embodiments considered.

The invention will be better understood and other characteristics, objects and advantages thereof will appear during the detailed description which follows and with reference to the appended drawings given by way of non-limiting examples and in which:

There already cited represents the structure of a resonator of an inertial sensor to which the invention applies.

There illustrates an example of an axisymmetric resonator having a plurality of transducers on the internal mass.

There illustrates an example of an axisymmetric resonator having a plurality of transducers on the two masses, internal and external.

There illustrates the frequency correction and quadrature bias correction controls according to the state of the art.

There illustrates the process according to the invention.

There illustrates a variant of the method according to the invention.

There illustrates a preferred embodiment of implementing steps B and C of the method according to the invention.

There illustrates an inertial sensor according to the invention.

There illustrates an embodiment of the inertial sensor according to the invention of the in gyrometer operation.

There illustrates an embodiment of the inertial sensor according to the invention of the in gyro operation.

There illustrates an embodiment of the inertial sensor according to the invention of the during the second phase of mixed mode.

For the sake of clarity, the same elements will bear the same references in the various figures.

DETAILED DESCRIPTION OF THE INVENTION

An in-depth analysis of the operation of the sensor shows that the difficulty in implementing frequency corrections and quadrature bias has its source in the fact that the TQ and TF transducers are positioned and operate along the X and Y sensor axes and not along the wave axes X'Y'. When the frequency and quadrature trim controls are applied as such to a wave vibrating at an electrical angle other than zero, the equalization and quadrature stiffnesses delivered correspond to the values that should be applied with transducers positioned and operating along the X'Y' axes of the wave marker. However, these transducers are fixed and positioned along the X and Y sensor axes. The stiffness values delivered by the first and the second servo-control are therefore not optimal for the vibration along X'. The objective of the method according to the invention is to deliver effective commands CTq and CTf for TQ and TF, that is to say adapted to the value of the electrical angle whatever its value.

In addition, the frequency differences between the two modes as well as the quadrature bias vary with the electrical angle of the vibration and the correction to be made also varies according to the angle due to the non-linearities of the sensor.

Finally, when the angle changes over time, the defects depending on the angle, it is no longer possible to filter over long times because the dynamics of the error is fast.

To solve this problem, the invention relates to a method 100 for determining a quadrature command CTq and a frequency command CTf of a vibration wave generated by a resonator Res of an inertial angular sensor, and the method s 'applying when the inertial sensor is in operation with a vibration wave vibrating along an axis X' characterized by an electrical angle θ whose different stages are illustrated .

The inertial sensor to which the invention applies comprises a resonator as described in the state of the art having a planar and axisymmetric structure around two axes X and Y perpendicular to each other defining an XY sensor reference and comprising two mobile masses vibrating M1 and M2 arranged one around the other, coupled together by coupling springs and configured to vibrate in tuning fork mode and in phase opposition along a direction X' defining a wave mark X'Y'.

The resonator comprises a plurality of electrostatic transducers controlled by electrical voltages and operating along the two axes X and Y, including at least, on one of the two masses:

- a pair of excitation transducers called E transducers configured to hold the wave at a constant amplitude via an amplitude command, to hold the wave flat and, if necessary, to rotate said vibration wave via a command of precession,
-a pair of detection transducers called transducers D configured to detect the movements of the vibration wave,
- a pair of quadrature bias compensation transducers called TQ transducers, configured to apply a quadrature stiffness via a quadrature command CTq, the quadrature stiffness being configured to cancel a coupling stiffness between X' and Y',
-a pair of frequency adjustment transducers called TF transducers, configured to apply an equalization stiffness via a CTf frequency command, the equalization stiffness being configured to cancel a stiffness difference between X' and Y' in such a way to equalize the resonance frequencies of the vibration wave on X' and Y'.

In a first step A, the electrical angle θ is determined.

In a second step B, first values Kq' and ΔK' of the quadrature and equalization stiffnesses are estimated from a first servo-control TrimQ and a second servo-control TrimF, respectively. These values are determined by drives of the same type as those operating according to the state of the art (apart from a few differences which will be detailed later) which "ignore" that the vibration wave vibrates along an axis X' different from X These values Kq' and ΔK', referred to as first values, are considered estimated in the reference wave X'Y' since they correspond to the values delivered by the automatic controls while the wave vibrates along X'. They correspond to the values that should be applied to TQ and TF with TQ and TF operating along X' and Y'.

In a following step C, second values Kq, ΔK of the quadrature and equalization stiffnesses in the sensor frame XY are determined, from the first values of said stiffnesses Kq', ΔK' estimated in step B. These values Kq and ΔK, called second values, are adapted to the fact that the transducers TQ and TF operate along the XY axes. Thus Kq' andΔK' are transformed into Kq and ΔK in such a way as to take into account the fact that the transducers applying the equalization and quadrature stiffnesses operate in the XY sensor frame and not in the X'Y' wave frame. In other words, from the values Kq' andΔK' estimated in the wave reference X'Y', a transformation is carried out on these two terms to return to the sensor reference XY in which the trimmers operate.

In a step D, the quadrature command CTq and the frequency command CTf corresponding respectively to the second values Kq and ΔK determined in step C are determined, in a conventional manner, and finally in a step E the frequency command CTf is applied and the quadrature command CTq determined in step D.

Thus the method according to the invention is applicable to a sensor with trim servos of the same type as those of the state of the art, that is to say without having to develop new servos, with a difference in signal processing level such as a transformation of the stiffnesses delivered by these automatic controls, carried out in step C of the method. Due to the accuracy of the stiffnesses Kq and ΔK calculated by the transformation, the force making the plane wave to be applied to E becomes zero again in steady state, the errors linked to the application of this force are eliminated, and all the correction (frequency and quadrature) is carried out via the TQ and TF transducers.

Due to this adaptation, the frequency and quadrature errors are eliminated and the inertial sensor delivers a measurement (speed or angle) which is no longer sensitive to the errors of the stiffness matrix whatever the value of the angle of θ vibration of the wave. It is recalled that the method according to the invention is used permanently and in parallel with the delivery of the measurement of the angular velocity or the angle of rotation along the sensitive axis Z.

Thanks to the implementation of the method according to the invention, the frequency deviations initially of about 3Hz are reduced to a few mHz, and quadrature errors of about 100°/s are reduced to less than 0.1°/s. These values, coupled with 100ppm class electronics in terms of phase errors, lead to drifts of less than one degree per hour.

According to a variant, the method is implemented while the inertial sensor operates in gyrometer mode. The electrical angle θ determined in step A is here equal to an angle θimp imposed on the vibration via the precession command. Different values of θimp can be used to average the errors, for example by taking measurements for θ equal to 30° then 60° then 90°. This implementation is made precise and efficient on a MEMs sensor thanks to the method according to the invention.

According to another variant, the method is implemented while the inertial sensor operates in gyroscope mode. The electrical angle θ then results from a rotation of the inertial sensor and is measured by it. The electrical angle determined in step A is equal to the measured angle of rotation θm.

According to yet another variant, the method according to the invention is implemented according to a mixed mode illustrated .

The method includes a first phase in which the electrical angle describes a plurality of electrical angles θi, i index, obtained by applying the precession command Cp. Steps A to E are implemented successively for each electrical angle θi. Step D comprises, in addition to the determination of CTf and Ctq, a sub-step MEM for storing the frequency command value CTfi associated with each angle θi and a sub-step MOD for determining a law of variation CTf (θi) of the frequency command as a function of the electrical angle. The form of this law is typically of the type Σ(akcos2kθ +bksin2kθ k typically varying from 0 to 4 and its determination consists in calculating the values of the coefficients ak and bk from a smoothing carried out on the measurement points θi.

This first phase can be implemented regardless of the operating mode of the sensor, gyrometer or gyroscope. It is preferably implemented in gyrometer mode. When the sensor operates in gyroscope mode (typically when the vehicle on which the sensor is mounted is stopped), the angle θi is obtained by sending a precession instruction to position itself at this angle from the known current angle. The first phase thus makes it possible to have a modeling of the CTf command to be applied according to the value of the electrical angle.

In a second phase, the inertial sensor operates in gyroscope mode, the electrical angle is left free and results from a rotation of the inertial sensor, which measures the value θm.

The second phase first of all comprises a step BO for putting the second TrimF servo-control into open loop. At this moment, the TrimF servo-control stops delivering the CTf command in real time, which is replaced by a CTf(θm) command determined from said variation law for the measured rotation angle θm. The command is applied over time by following the variations of θm. At the same time, the resonance frequency difference Δf between the two axes of the wave is measured. The principle of frequency trimming is to send a disturbance, measure the Δf and then correct it via CTf. When in open loop it is possible to continue to send the disturbance, measure the Δf but the CTf correction is no longer applied.

The open looping step BO is implemented as long as the resonance frequency difference is less than or equal to a predetermined threshold S. The difference Δf evolves according to the temperature and with aging. When the frequency difference Δf becomes greater than the threshold, the second TrimF servo-control is returned to closed loop to allow the variation law to be updated by starting again on the first phase. Throughout the second phase, the first servo-control continues to operate as in the first phase. This mixed mode has many advantages.

. Frequency trimming injects noise into the angular velocity measurement when the servo-control is in closed loop (hence the interest of open-loop operation). The non-intrusive quadrature trimming remains in a closed loop.

The frequency differences between the two modes vary with the angle. This is not just a geometric problem, in which case it would be sufficient to simply correct the frequency deviation at zero angle and the correction would be valid for any angle after rotation. There are also frequency deviations related to non-linearities which cause the correction to change depending on the angle.

When you are in gyroscope mode, you let the wave spin. So potentially, the angle of the vibration can change quickly, and therefore the frequency deviation (to be corrected) can change quickly, and therefore the TrimF frequency servo must have a bandwidth as high as the maximum angular speed. This poses a problem because for this servo-control the signal-to-noise ratio is very low and it is necessary to filter over a long time to achieve efficient servo-control. As a result, the frequency deviation cannot be filtered over a long time, generating significant frequency deviations in the event of noise: Hence the interest of being able to perform frequency trimming at different angles in gyrometer mode. Since the angle is constant in gyro mode, filtering may take longer. Once the frequency deviations have been identified, they can be corrected in open loop and return to gyroscope mode. The appropriate command is then directly applied by following a rapid variation of ΔK which is therefore compatible with a high dynamic. Thus the gyroscope mode benefits from the results of the frequency trimming in gyrometer mode. Note that for low angular velocities, frequency trimming in gyroscope mode can still be implemented.

Preferably steps B and C are implemented according to embodiments established from the results of a matrix calculation developed below, the method according to the invention according to this embodiment being illustrated .

Step B of determining the torque (Kq', ΔK') by the signal processing of the TrimF and TrimQ servo-controls is carried out by having previously carried out a change of frame, in the form of a sub-step B1 of determining the position (x', y') of the vibration wave in the X'Y' wave frame from the measurement of the position (x, y) of the vibration wave in the XY sensor frame made by the transducers D and the electrical angle θWe therefore have:

x' = cosθ.x + sinθ.y

y’ = -sinθ.x + cosθ.y

The pair (x',y') is thus used as input to processing.

In a sub-step B2, the estimation of the first values (Kq', ΔK') of the quadrature and equalization stiffnesses is carried out from the position in the wave frame (x', y').

Step C consists in determining a vector defined by the second values having as coordinates (Kq, ΔK) by applying a rotation of an angle equal to twice the electrical angle, i.e. 2θ to the vector having as coordinates the first values (Kq', ΔK').

The transformation of the pair (Kq', ΔK') into (Kq, ΔK) is therefore expressed by the matrix relation (the coordinates (Kq, ΔK) are expressed in the XY frame):

Let: Kq = cos2θ.Kq' + sin2θ.ΔK' and ΔKq = -sin2θ.Kq' + cos2θ.ΔK'

These transformation relationships are programmed into the servos. In theory, once returned to the sensor reference, the Kq and ΔK values are constant and valid for all theθ. But due to non-linearities, these values depend on θ and are variable over time due to temperature and sensor aging, so they should be recalculated in real time.

We are now going to demonstrate how the relation Math1 is obtained.

One starts from the matrix of true stiffness K' expressed in the reference wave X'Y'. We will use the term “‘” for quantities expressed in the wave reference and the term without “‘” for quantities expressed in the XY sensor reference.

We correct this stiffness matrix using the “trimming” combs with the following correction matrix Kc’:

The final stiffness matrix Kf' is equal to:

For a perfect correction we have:

Once ΔK' and Kq' have been determined in reference X'Y' (step B) it is necessary to determine ΔK and Kq in reference XY (step C).

We start from the correction matrix Kc', which is a linear application that transforms a vector Ve' expressed in X'Y' into a vector Vs' also expressed in reference X'Y': Vs' = Kc' Ve'

We wish to determine the same transformation that we will note Kc allowing to pass from a vector Ve in XY reference to a vector Vs also expressed in XY reference.

We note R(θ) the rotation which makes it possible to pass from the XY reference to the X'Y' reference, V a vector expressed in the XY reference and V' is the same vector expressed in the X'Y' reference. We have :

and Vs’ = Kc’.Ve’ i.e.:

And so :

Either :

We can also write Kc as a function of ΔK and Kq:

We can then identify the terms:

Which give :

According to another aspect, the invention relates to an inertial angular sensor 10 illustrated comprising a Res resonator as described previously comprising a plurality of electrostatic transducers controlled by electrical voltages and operating along the two axes X and Y, including at least, on one of the two masses: a pair of excitation transducers generically named transducers E, a pair of sensing transducers generically referred to as D transducers, a pair of quadrature bias compensation transducers generically referred to as TQ transducers, and a pair of frequency adjustment transducers generically referred to as TF transducers .

The quadrature stiffness is determined by a first TrimQ servo and the equalization stiffness is determined by a second TrimF servo. The sensor further comprises a processing unit UT. The UT unit is configured to determine the electrical angle θof the vibration.

The processing unit UT comprises a first module 20 configured to estimate first values Kq' and ΔK' of the quadrature and equalization stiffnesses from the first servo-control TrimQ and from the second servo-control TrimF respectively, the first values being estimated in the X'Y' wave marker.

The processing unit also comprises a second module 21 configured to determine second values Kq and ΔK of the quadrature and equalization stiffnesses in the XY sensor frame, from the first values of the stiffnesses Kq' and ΔK'.

The unit UT also comprises a set of two electrical gain modules Gq and Gf configured to respectively determine the quadrature command CTq corresponding to the second value of the quadrature stiffness Kq and the frequency command CTf corresponding to said second stiffness value equalization ΔK, the transducers TF and TQ being configured to respectively apply the frequency command CTf and the quadrature command CTq.

According to a preferred embodiment, the first module 20 is configured to determine the position (x', y') of the vibration wave in the wave reference X'Y' from the electrical angle θ and the measurement of the position (x, y) of the vibration wave in the XY sensor reference produced by the transducers D, and to estimate first values Kq' and ΔK' of the quadrature and equalization stiffnesses from said position in the wave marker (x', y'). According to a preferred mode, the second module 21 is configured to determine a vector (Kq, ΔK) whose coordinates are the second values by applying a rotation by an angle equal to twice the electrical angle 2θ to the vector (Kq' , ΔK') whose coordinates are the first values.

There illustrates an embodiment of the inertial sensor 10 according to the invention of the in gyrometer operation. The angle θ is imposed by the precession command Cp applied to the transducer E. The block 20 comprises a module 2 to effect the change of frame, from the sensor frame XY to the wave frame X'Y'. Module 2 performs a rotation R1(θ). From (x', y') the determination of Kq' and ΔK' is carried out by processing using elements of the processing according to the state of the art: block 3 is equivalent to Coq1, block 5 to Coq2, block 4 to Cof and block 6 to Coq3. The module 21 performs a rotation R2(2θ) as explained above.

There illustrates an embodiment of the inertial sensor 10 according to the invention of the in gyro operation. The angle θm is left free and measured. The measurement of θm is carried out in a conventional manner via a module 7 delivering the speed of rotation Ω and an integrator 8.

There illustrates an embodiment of the inertial sensor 10 according to the invention of the during the second phase of mixed mode. The TrimF servo-control is put in open loop, and the command applied to the transducer TF comes from the module 9 in which the law of variation CTf(θ) is stored, applied to the angle θm.

Claims (9)

Method (100) for determining a quadrature command (CTq) and a frequency command (CTf) of a vibration wave generated by a resonator (Res) of an inertial angular sensor (10),
- the resonator (Res) having a plane and axisymmetric structure around two mutually perpendicular X and Y axes defining an XY sensor marker and comprising two vibrating mobile masses (M1, M2) arranged one around the other, coupled between they by coupling springs and configured to vibrate in phase opposition along a direction X' defining a wave marker X'Y',
- the resonator further comprising a plurality of electrostatic transducers controlled by electrical voltages and operating along the two axes X and Y, including at least, on at least one of the two masses:
- a pair of excitation transducers called E transducers configured to maintain the wave at a constant amplitude via an amplitude command (Ca) and, if necessary, to rotate said vibration wave via a precession command (Cp ),
-a pair of detection transducers called transducers D configured to detect the movements of the vibration wave,
- a pair of quadrature bias compensation transducers referred to as TQ transducers, configured to apply quadrature stiffness via quadrature control (CTq), the quadrature stiffness being configured to cancel a coupling stiffness between X' and Y ',
-a pair of frequency adjustment transducers called TF transducers, configured to apply an equalization stiffness via a frequency command (CTf), the equalization stiffness being configured to cancel a stiffness difference between X' and Y' so as to equalize the resonance frequencies of the vibration wave on X' and Y',
the method applying when the inertial sensor is in operation with a vibration wave vibrating along X' characterized by an electrical angle (θ),
the method comprising the steps of:
-To determine the electrical angle (θ),
-B estimate first values (Kq', ΔK') of said quadrature and equalization stiffnesses from respectively a first (TrimQ) and a second (TrimF) servo-control, said first values being estimated in the wave frame X'Y',
-C determining second values (Kq, ΔK) of said quadrature and equalization stiffnesses in the XY sensor frame, from the first values of said stiffnesses (Kq',ΔK') estimated in step B,
-D determining the quadrature command (CTq) and the frequency command (CTf) corresponding respectively to said second values (Kq, ΔK) determined in step C,
-E apply the frequency command (CTf) and quadrature command (CTq) determined in step D. Method according to Claim 1, in which the inertial sensor operates in gyrometer mode, the electrical angle (θ) determined in step A being equal to an imposed angle (θimp) via the precession command (Cp). Method according to claim 1, in which the inertial sensor operates in gyroscope mode, the electrical angle (θ) resulting from a rotation of the inertial sensor being measured by said inertial sensor, the electrical angle determined in step A being equal to said measured angle of rotation (θm). A method (100) according to claim 1 comprising:
- a first phase in which the electrical angle describes a plurality of electrical angles (θi) obtained by applying said precession command (Cp), steps A to E being implemented for each electrical angle (θi), the step D further comprising a sub-step (MEM) for storing the associated frequency command value (CTfi) and a sub-step (MOD) for determining a law of variation (CTf(θ)) of the frequency command depending on the electrical angle,
- a second phase in which the inertial sensor operates in gyroscope mode, the electrical angle (θ) left free resulting from a rotation of the inertial sensor and being measured (θm) by said inertial sensor, the second phase comprising:
*a step (BO) of open-looping the second servo-control (TrimF), the frequency command applied then being determined from said law of variation for said measured angle of rotation (θm),
*a step of detecting a resonance frequency deviation (Δf), the open looping step (BO) being implemented as long as said resonance frequency deviation is less than or equal to a predetermined threshold (S ),
*a closed-loop reset step of said second servo-control (TrimF) when the frequency deviation (Δf) is greater than said threshold, the method then looping back to the first phase for an update of said variation law. Method according to one of the preceding claims, in which step B comprises:
-a sub-step B1 for determining a position (x', y') of the vibration wave in the wave reference X'Y' from the measurement of a position (x, y) of the wave of vibration in the XY sensor frame and of the electrical angle (θ),
a sub-step B2 of estimating the first values (Kq', ΔK') of said quadrature and equalization stiffnesses from said position in the wave frame (x', y'). Method according to one of the preceding claims, in which step C consists in determining a vector (Kq, ΔK) defined by said second values by applying a rotation by an angle equal to twice the electrical angle (2θ) to the vector (Kq', ΔK') defined by said first values. Inertial angular sensor (10) comprising:
-a resonator (Res) having a planar and axisymmetric structure around two mutually perpendicular X and Y axes defining an XY sensor marker, and comprising two vibrating mobile masses (M1, M2) arranged one around the other, coupled between them by coupling springs and configured to vibrate in phase opposition according to a vibration wave (OV) vibrating in a direction X' characterized by an electrical angle (θ) and defining a wave mark X'Y',
- the resonator further comprising a plurality of electrostatic transducers controlled by electrical voltages and operating along the two axes X and Y, including at least, on at least one of the two masses:
- a pair of excitation transducers called E transducers configured to maintain the wave at a constant amplitude via an amplitude command (Ca) and, if necessary, to rotate said vibration wave via a precession command (Cp ),
-a pair of detection transducers called transducers D configured to detect the movements of the vibration wave,
- a pair of quadrature bias compensation transducers referred to as TQ transducers, configured to apply quadrature stiffness via quadrature control (CTq), the quadrature stiffness being configured to cancel a coupling stiffness between X' and Y ',
-a pair of frequency adjustment transducers called TF transducers, configured to apply an equalization stiffness via a frequency command (CTf), the equalization stiffness being configured to cancel a stiffness difference between X' and Y' so as to equalize the resonance frequencies of the vibration wave on X' and Y',
- said quadrature and equalization stiffnesses being determined from respectively a first (TrimQ) and a second (TrimF) servo-control,
-the sensor further comprising a processing unit (UT) configured to determine said electrical angle (θ) and comprising:
-a first module (20) configured to estimate first values (Kq', ΔK') of said quadrature and equalization stiffnesses from respectively the first (TrimQ) and the second (TrimF) slaving, said first values being estimated in the X'Y' wave marker,
-a second module (21) configured to determine second values (Kq, ΔK) of said quadrature and equalization stiffnesses in the XY sensor frame, from the first values of said stiffnesses (Kq',ΔK'),
-a set of two electrical gain modules (Gq, Gf) configured to respectively determine the quadrature command (CTq) corresponding to the second value of the quadrature stiffness (Kq) and the frequency command (CTf) corresponding to said second equalization stiffness value (ΔK),
said transducers TF and TQ being configured to respectively apply said frequency command (CTf) and said quadrature command (CTq) to the resonator. Inertial sensor according to Claim 7, in which the first module (20) is configured to determine a position (x', y') of the vibration wave in the wave frame X'Y' from the electrical angle (θ ) and the measurement of a position (x, y) of the vibration wave in the sensor frame XY carried out by the transducers D, and to estimate first values (Kq', ΔK') of said quadrature stiffnesses and d 'equalization from the said position in the wave reference (x', y'). Inertial sensor according to one of Claims 7 or 8, in which the second module (21) is configured to determine a vector (Kq, ΔK) defined by the said second values by applying a rotation through an angle equal to twice l 'electric angle (2θ) to the vector (Kq', ΔK') defined by said first values.